chore: sync VERSION file with release v1.1.261 [skip ci]

- Fix updateApiKey to use JSON.stringify for permissions field
- Add comma-separated string handling in normalizePermissions
- Add frontend parsing for comma-separated permissions format

Fixes issue where selecting multiple permissions (e.g. Claude + OpenAI)
would be saved as "claude,openai" instead of '["claude","openai"]'

.env.example

@@ -33,6 +33,59 @@ CLAUDE_API_URL=https://api.anthropic.com/v1/messages
 CLAUDE_API_VERSION=2023-06-01
 CLAUDE_BETA_HEADER=claude-code-20250219,oauth-2025-04-20,interleaved-thinking-2025-05-14,fine-grained-tool-streaming-2025-05-14
 
+# 🤖 Gemini OAuth / Antigravity 配置（可选）
+# 不配置时使用内置默认值；如需自定义或避免在代码中出现 client secret，可在此覆盖
+# GEMINI_OAUTH_CLIENT_ID=
+# GEMINI_OAUTH_CLIENT_SECRET=
+# Gemini CLI OAuth redirect_uri（可选，默认 https://codeassist.google.com/authcode）
+# GEMINI_OAUTH_REDIRECT_URI=
+# ANTIGRAVITY_OAUTH_CLIENT_ID=
+# ANTIGRAVITY_OAUTH_CLIENT_SECRET=
+# Antigravity OAuth redirect_uri（可选，默认 http://localhost:45462；用于避免 redirect_uri_mismatch）
+# ANTIGRAVITY_OAUTH_REDIRECT_URI=http://localhost:45462
+# Antigravity 上游地址（可选，默认 sandbox）
+# ANTIGRAVITY_API_URL=https://daily-cloudcode-pa.sandbox.googleapis.com
+# Antigravity User-Agent（可选）
+# ANTIGRAVITY_USER_AGENT=antigravity/1.11.3 windows/amd64
+
+# Claude Code（Anthropic Messages API）路由分流（无需额外环境变量）：
+# - /api             -> Claude 账号池（默认）
+# - /antigravity/api  -> Antigravity OAuth
+# - /gemini-cli/api   -> Gemini CLI OAuth
+
+# ============================================================================
+# 🐛 调试 Dump 配置（可选）
+# ============================================================================
+# 以下开启后会在项目根目录写入 .jsonl 调试文件，便于排查问题。
+# ⚠️ 生产环境建议关闭，避免磁盘占用。
+#
+# 📄 输出文件列表：
+#   - anthropic-requests-dump.jsonl      (客户端请求)
+#   - anthropic-responses-dump.jsonl     (返回给客户端的响应)
+#   - anthropic-tools-dump.jsonl         (工具定义快照)
+#   - antigravity-upstream-requests-dump.jsonl   (发往上游的请求)
+#   - antigravity-upstream-responses-dump.jsonl  (上游 SSE 响应)
+#
+# 📌 开关配置：
+# ANTHROPIC_DEBUG_REQUEST_DUMP=true
+# ANTHROPIC_DEBUG_RESPONSE_DUMP=true
+# ANTHROPIC_DEBUG_TOOLS_DUMP=true
+# ANTIGRAVITY_DEBUG_UPSTREAM_REQUEST_DUMP=true
+# ANTIGRAVITY_DEBUG_UPSTREAM_RESPONSE_DUMP=true
+#
+# 📏 单条记录大小上限（字节），默认 2MB：
+# ANTHROPIC_DEBUG_REQUEST_DUMP_MAX_BYTES=2097152
+# ANTHROPIC_DEBUG_RESPONSE_DUMP_MAX_BYTES=2097152
+# ANTIGRAVITY_DEBUG_UPSTREAM_REQUEST_DUMP_MAX_BYTES=2097152
+#
+# 📦 整个 Dump 文件大小上限（字节），超过后自动轮转为 .bak 文件，默认 10MB：
+# DUMP_MAX_FILE_SIZE_BYTES=10485760
+#
+# 🔧 工具失败继续：当 tool_result 标记 is_error=true 时，提示模型不要中断任务
+# （仅 /antigravity/api 分流生效）
+# ANTHROPIC_TOOL_ERROR_CONTINUE=true
+
+
 # 🚫 529错误处理配置
 # 启用529错误处理，0表示禁用，>0表示过载状态持续时间（分钟）
 CLAUDE_OVERLOAD_HANDLING_MINUTES=0
@@ -61,6 +114,9 @@ PROXY_USE_IPV4=true
 # ⏱️ 请求超时配置
 REQUEST_TIMEOUT=600000  # 请求超时设置（毫秒），默认10分钟
 
+# 🔧 请求体大小配置
+REQUEST_MAX_SIZE_MB=60
+
 # 📈 使用限制
 DEFAULT_TOKEN_LIMIT=1000000
 
@@ -75,6 +131,8 @@ TOKEN_USAGE_RETENTION=2592000000
 HEALTH_CHECK_INTERVAL=60000
 TIMEZONE_OFFSET=8  # UTC偏移小时数，默认+8（中国时区）
 METRICS_WINDOW=5  # 实时指标统计窗口（分钟），可选1-60，默认5分钟
+# 启动时清理残留的并发排队计数器（默认true，多实例部署时建议设为false）
+CLEAR_CONCURRENCY_QUEUES_ON_STARTUP=true
 
 # 🎨 Web 界面配置
 WEB_TITLE=Claude Relay Service

.gitignore

@@ -26,6 +26,7 @@ redis_data/
 
 # Logs directory
 logs/
+logs1/
 *.log
 startup.log
 app.log

CLAUDE.md

@@ -22,6 +22,7 @@ Claude Relay Service 是一个多平台 AI API 中转服务，支持 **Claude (
 - **权限控制**: API Key支持权限配置（all/claude/gemini/openai等），控制可访问的服务类型
 - **客户端限制**: 基于User-Agent的客户端识别和限制，支持ClaudeCode、Gemini-CLI等预定义客户端
 - **模型黑名单**: 支持API Key级别的模型访问限制
+- **并发请求排队**: 当API Key并发数超限时，请求进入队列等待而非立即返回429，支持配置最大排队数、超时时间，适用于Claude Code Agent并行工具调用场景
 
 ### 主要服务组件
 
@@ -60,6 +61,7 @@ Claude Relay Service 是一个多平台 AI API 中转服务，支持 **Claude (
 
 - **apiKeyService.js**: API Key管理，验证、限流、使用统计、成本计算
 - **userService.js**: 用户管理系统，支持用户注册、登录、API Key管理
+- **userMessageQueueService.js**: 用户消息串行队列，防止同账户并发用户消息触发限流
 - **pricingService.js**: 定价服务，模型价格管理和成本计算
 - **costInitService.js**: 成本数据初始化服务
 - **webhookService.js**: Webhook通知服务
@@ -185,12 +187,17 @@ npm run service:stop          # 停止服务
 - `CLAUDE_OVERLOAD_HANDLING_MINUTES`: Claude 529错误处理持续时间（分钟，0表示禁用）
 - `STICKY_SESSION_TTL_HOURS`: 粘性会话TTL（小时，默认1）
 - `STICKY_SESSION_RENEWAL_THRESHOLD_MINUTES`: 粘性会话续期阈值（分钟，默认0）
+- `USER_MESSAGE_QUEUE_ENABLED`: 启用用户消息串行队列（默认false）
+- `USER_MESSAGE_QUEUE_DELAY_MS`: 用户消息请求间隔（毫秒，默认200）
+- `USER_MESSAGE_QUEUE_TIMEOUT_MS`: 队列等待超时（毫秒，默认5000，锁持有时间短无需长等待）
+- `USER_MESSAGE_QUEUE_LOCK_TTL_MS`: 锁TTL（毫秒，默认5000，请求发送后立即释放无需长TTL）
 - `METRICS_WINDOW`: 实时指标统计窗口（分钟，1-60，默认5）
 - `MAX_API_KEYS_PER_USER`: 每用户最大API Key数量（默认1）
 - `ALLOW_USER_DELETE_API_KEYS`: 允许用户删除自己的API Keys（默认false）
 - `DEBUG_HTTP_TRAFFIC`: 启用HTTP请求/响应调试日志（默认false，仅开发环境）
 - `PROXY_USE_IPV4`: 代理使用IPv4（默认true）
 - `REQUEST_TIMEOUT`: 请求超时时间（毫秒，默认600000即10分钟）
+- `CLEAR_CONCURRENCY_QUEUES_ON_STARTUP`: 启动时清理残留的并发排队计数器（默认true，多实例部署时建议设为false）
 
 #### AWS Bedrock配置（可选）
 - `CLAUDE_CODE_USE_BEDROCK`: 启用Bedrock（设置为1启用）
@@ -337,6 +344,35 @@ npm run setup  # 自动生成密钥并创建管理员账户
 11. **速率限制未清理**: rateLimitCleanupService每5分钟自动清理过期限流状态
 12. **成本统计不准确**: 运行 `npm run init:costs` 初始化成本数据，检查pricingService是否正确加载模型价格
 13. **缓存命中率低**: 查看缓存监控统计，调整LRU缓存大小配置
+14. **用户消息队列超时**: 优化后锁持有时间已从分钟级降到毫秒级（请求发送后立即释放），默认 `USER_MESSAGE_QUEUE_TIMEOUT_MS=5000` 已足够。如仍有超时，检查网络延迟或禁用此功能（`USER_MESSAGE_QUEUE_ENABLED=false`）
+15. **并发请求排队问题**:
+   - 排队超时：检查 `concurrentRequestQueueTimeoutMs` 配置是否合理（默认10秒）
+   - 排队数过多：调整 `concurrentRequestQueueMaxSize` 和 `concurrentRequestQueueMaxSizeMultiplier`
+   - 查看排队统计：访问 `/admin/concurrency-queue/stats` 接口查看 entered/success/timeout/cancelled/socket_changed/rejected_overload 统计
+   - 排队计数泄漏：系统重启时自动清理，或访问 `/admin/concurrency-queue` DELETE 接口手动清理
+   - Socket 身份验证失败：查看 `socket_changed` 统计，如果频繁发生，检查代理配置或客户端连接稳定性
+   - 健康检查拒绝：查看 `rejected_overload` 统计，表示队列过载时的快速失败次数
+
+### 代理配置要求（并发请求排队）
+
+使用并发请求排队功能时，需要正确配置代理（如 Nginx）的超时参数：
+
+- **推荐配置**: `proxy_read_timeout >= max(2 × concurrentRequestQueueTimeoutMs, 60s)`
+  - 当前默认排队超时 10 秒，Nginx 默认 `proxy_read_timeout = 60s` 已满足要求
+  - 如果调整排队超时到 60 秒，推荐代理超时 ≥ 120 秒
+- **Nginx 配置示例**:
+  ```nginx
+  location /api/ {
+      proxy_read_timeout 120s;  # 排队超时 60s 时推荐 120s
+      proxy_connect_timeout 10s;
+      # ...其他配置
+  }
+  ```
+- **企业防火墙环境**:
+  - 某些企业防火墙可能静默关闭长时间无数据的连接（20-40 秒）
+  - 如遇此问题，联系网络管理员调整空闲连接超时策略
+  - 或降低 `concurrentRequestQueueTimeoutMs` 配置
+- **后续升级说明**: 如有需要，后续版本可能提供可选的轻量级心跳机制
 
 ### 调试工具
 
@@ -449,6 +485,15 @@ npm run setup  # 自动生成密钥并创建管理员账户
 - **缓存优化**: 多层LRU缓存（解密缓存、账户缓存），全局缓存监控和统计
 - **成本追踪**: 实时token使用统计（input/output/cache_create/cache_read）和成本计算（基于pricingService）
 - **并发控制**: Redis Sorted Set实现的并发计数，支持自动过期清理
+- **并发请求排队**: 当API Key并发超限时，请求进入队列等待而非直接返回429
+  - **工作原理**: 采用「先占后检查」模式，每次轮询尝试占位，超限则释放继续等待
+  - **指数退避**: 初始200ms，指数增长至最大2秒，带±20%抖动防惊群效应
+  - **智能清理**: 排队计数有TTL保护（超时+30秒），进程崩溃也能自动清理
+  - **Socket身份验证**: 使用UUID token + socket对象引用双重验证，避免HTTP Keep-Alive连接复用导致的身份混淆
+  - **健康检查**: P90等待时间超过阈值时快速失败（返回429），避免新请求在过载时继续排队
+  - **配置参数**: `concurrentRequestQueueEnabled`（默认false）、`concurrentRequestQueueMaxSize`（默认3）、`concurrentRequestQueueMaxSizeMultiplier`（默认0）、`concurrentRequestQueueTimeoutMs`（默认10秒）、`concurrentRequestQueueMaxRedisFailCount`（默认5）、`concurrentRequestQueueHealthCheckEnabled`（默认true）、`concurrentRequestQueueHealthThreshold`（默认0.8）
+  - **最大排队数**: max(固定值, 并发限制×倍数)，例如并发限制=10、倍数=2时最大排队数=20
+  - **适用场景**: Claude Code Agent并行工具调用、批量请求处理
 - **客户端识别**: 基于User-Agent的客户端限制，支持预定义客户端（ClaudeCode、Gemini-CLI等）
 - **错误处理**: 529错误自动标记账户过载状态，配置时长内自动排除该账户
 
@@ -508,8 +553,16 @@ npm run setup  # 自动生成密钥并创建管理员账户
   - `overload:{accountId}` - 账户过载状态（529错误）
 - **并发控制**:
   - `concurrency:{accountId}` - Redis Sorted Set实现的并发计数
+- **并发请求排队**:
+  - `concurrency:queue:{apiKeyId}` - API Key级别的排队计数器（TTL由 `concurrentRequestQueueTimeoutMs` + 30秒缓冲决定）
+  - `concurrency:queue:stats:{apiKeyId}` - 排队统计（entered/success/timeout/cancelled）
+  - `concurrency:queue:wait_times:{apiKeyId}` - 按API Key的等待时间记录（用于P50/P90/P99计算）
+  - `concurrency:queue:wait_times:global` - 全局等待时间记录
 - **Webhook配置**:
   - `webhook_config:{id}` - Webhook配置
+- **用户消息队列**:
+  - `user_msg_queue_lock:{accountId}` - 用户消息队列锁（当前持有者requestId）
+  - `user_msg_queue_last:{accountId}` - 上次请求完成时间戳（用于延迟计算）
 - **系统信息**:
   - `system_info` - 系统状态缓存
   - `model_pricing` - 模型价格数据（pricingService）

Dockerfile

@@ -1,4 +1,17 @@
-# 🎯 前端构建阶段
+# 🎯 后端依赖阶段 (与前端构建并行)
+FROM node:18-alpine AS backend-deps
+
+# 📁 设置工作目录
+WORKDIR /app
+
+# 📦 复制 package 文件
+COPY package*.json ./
+
+# 🔽 安装依赖 (生产环境) - 使用 BuildKit 缓存加速
+RUN --mount=type=cache,target=/root/.npm \
+    npm ci --only=production
+
+# 🎯 前端构建阶段 (与后端依赖并行)
 FROM node:18-alpine AS frontend-builder
 
 # 📁 设置工作目录
@@ -7,8 +20,9 @@ WORKDIR /app/web/admin-spa
 # 📦 复制前端依赖文件
 COPY web/admin-spa/package*.json ./
 
-# 🔽 安装前端依赖
-RUN npm ci
+# 🔽 安装前端依赖 - 使用 BuildKit 缓存加速
+RUN --mount=type=cache,target=/root/.npm \
+    npm ci
 
 # 📋 复制前端源代码
 COPY web/admin-spa/ ./
@@ -34,17 +48,16 @@ RUN apk add --no-cache \
 # 📁 设置工作目录
 WORKDIR /app
 
-# 📦 复制 package 文件
+# 📦 复制 package 文件 (用于版本信息等)
 COPY package*.json ./
 
-# 🔽 安装依赖 (生产环境)
-RUN npm ci --only=production && \
-    npm cache clean --force
+# 📦 从后端依赖阶段复制 node_modules (已预装好)
+COPY --from=backend-deps /app/node_modules ./node_modules
 
 # 📋 复制应用代码
 COPY . .
 
-# 📦 从构建阶段复制前端产物
+# 📦 从前端构建阶段复制前端产物
 COPY --from=frontend-builder /app/web/admin-spa/dist /app/web/admin-spa/dist
 
 # 🔧 复制并设置启动脚本权限

README.md

@@ -1,5 +1,10 @@
 # Claude Relay Service
 
+> [!CAUTION]
+> **安全更新通知**：v1.1.248 及以下版本存在严重的管理员认证绕过漏洞，攻击者可未授权访问管理面板。
+>
+> **请立即更新到 v1.1.249+ 版本**，或迁移到新一代项目 **[CRS 2.0 (sub2api)](https://github.com/Wei-Shaw/sub2api)**
+
 <div align="center">
 
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
@@ -389,6 +394,9 @@ docker-compose.yml 已包含：
 
 **Claude Code 设置环境变量：**
 
+
+**使用标准 Claude 账号池**
+
 默认使用标准 Claude 账号池：
 
 ```bash
@@ -396,6 +404,24 @@ export ANTHROPIC_BASE_URL="http://127.0.0.1:3000/api/" # 根据实际填写你
 export ANTHROPIC_AUTH_TOKEN="后台创建的API密钥"
 ```
 
+**使用 Antigravity 账户池**
+
+适用于通过 Antigravity 渠道使用 Claude 模型（如 `claude-opus-4-5` 等）。
+
+```bash
+# 1. 设置 Base URL 为 Antigravity 专用路径
+export ANTHROPIC_BASE_URL="http://127.0.0.1:3000/antigravity/api/"
+
+# 2. 设置 API Key（在后台创建，权限需包含 'all' 或 'gemini'）
+export ANTHROPIC_AUTH_TOKEN="后台创建的API密钥"
+
+# 3. 指定模型名称（直接使用短名，无需前缀！）
+export ANTHROPIC_MODEL="claude-opus-4-5"
+
+# 4. 启动
+claude
+```
+
 **VSCode Claude 插件配置：**
 
 如果使用 VSCode 的 Claude 插件，需要在 `~/.claude/config.json` 文件中配置：
@@ -408,6 +434,8 @@ export ANTHROPIC_AUTH_TOKEN="后台创建的API密钥"
 
 如果该文件不存在，请手动创建。Windows 用户路径为 `C:\Users\你的用户名\.claude\config.json`。
 
+> 💡 **IntelliJ IDEA 用户推荐**：[Claude Code Plus](https://github.com/touwaeriol/claude-code-plus) - 将 Claude Code 直接集成到 IDE，支持代码理解、文件读写、命令执行。插件市场搜索 `Claude Code Plus` 即可安装。
+
 **Gemini CLI 设置环境变量：**
 
 **方式一（推荐）：通过 Gemini Assist API 方式访问**
@@ -597,8 +625,9 @@ gpt-5                      # Codex使用固定模型ID
 - 所有账号类型都使用相同的API密钥（在后台统一创建）
 - 根据不同的路由前缀自动识别账号类型
 - `/claude/` - 使用Claude账号池
+- `/antigravity/api/` - 使用Antigravity账号池（推荐用于Claude Code）
 - `/droid/claude/` - 使用Droid类型Claude账号池（只建议api调用或Droid Cli中使用）
-- `/gemini/` - 使用Gemini账号池  
+- `/gemini/` - 使用Gemini账号池
 - `/openai/` - 使用Codex账号（只支持Openai-Response格式）
 - `/droid/openai/` - 使用Droid类型OpenAI兼容账号池（只建议api调用或Droid Cli中使用）
 - 支持所有标准API端点（messages、models等）

README_EN.md

@@ -1,5 +1,10 @@
 # Claude Relay Service
 
+> [!CAUTION]
+> **Security Update**: v1.1.248 and below contain a critical admin authentication bypass vulnerability allowing unauthorized access to the admin panel.
+>
+> **Please update to v1.1.249+ immediately**, or migrate to the next-generation project **[CRS 2.0 (sub2api)](https://github.com/Wei-Shaw/sub2api)**
+
 <div align="center">
 
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)

SECURITY.md

@@ -0,0 +1,21 @@
+# Security Policy
+
+## Supported Versions
+
+Use this section to tell people about which versions of your project are
+currently being supported with security updates.
+
+| Version | Supported          |
+| ------- | ------------------ |
+| 5.1.x   | :white_check_mark: |
+| 5.0.x   | :x:                |
+| 4.0.x   | :white_check_mark: |
+| < 4.0   | :x:                |
+
+## Reporting a Vulnerability
+
+Use this section to tell people how to report a vulnerability.
+
+Tell them where to go, how often they can expect to get an update on a
+reported vulnerability, what to expect if the vulnerability is accepted or
+declined, etc.

VERSION

@@ -1 +1 @@
-1.1.216
+1.1.261

config/config.example.js

@@ -203,6 +203,23 @@ const config = {
   development: {
     debug: process.env.DEBUG === 'true',
     hotReload: process.env.HOT_RELOAD === 'true'
+  },
+
+  // 💰 账户余额相关配置
+  accountBalance: {
+    // 是否允许执行自定义余额脚本（安全开关）
+    // 说明：脚本能力可发起任意 HTTP 请求并在服务端执行 extractor 逻辑，建议仅在受控环境开启
+    // 默认保持开启；如需禁用请显式设置：BALANCE_SCRIPT_ENABLED=false
+    enableBalanceScript: process.env.BALANCE_SCRIPT_ENABLED !== 'false'
+  },
+
+  // 📬 用户消息队列配置
+  // 优化说明：锁在请求发送成功后立即释放（而非请求完成后），因为 Claude API 限流基于请求发送时刻计算
+  userMessageQueue: {
+    enabled: process.env.USER_MESSAGE_QUEUE_ENABLED === 'true', // 默认关闭
+    delayMs: parseInt(process.env.USER_MESSAGE_QUEUE_DELAY_MS) || 200, // 请求间隔（毫秒）
+    timeoutMs: parseInt(process.env.USER_MESSAGE_QUEUE_TIMEOUT_MS) || 5000, // 队列等待超时（毫秒），锁持有时间短，无需长等待
+    lockTtlMs: parseInt(process.env.USER_MESSAGE_QUEUE_LOCK_TTL_MS) || 5000 // 锁TTL（毫秒），5秒足以覆盖请求发送
   }
 }
 

docker-compose.yml

@@ -21,6 +21,9 @@ services:
       - PORT=3000
       - HOST=0.0.0.0
       
+      # 🔧 请求体大小配置
+      - REQUEST_MAX_SIZE_MB=60
+      
       # 🔐 安全配置（必填）
       - JWT_SECRET=${JWT_SECRET}  # 必填：至少32字符的随机字符串
       - ENCRYPTION_KEY=${ENCRYPTION_KEY}  # 必填：32字符的加密密钥

package-lock.json

@@ -20,12 +20,14 @@
         "dotenv": "^16.3.1",
         "express": "^4.18.2",
         "google-auth-library": "^10.1.0",
+        "heapdump": "^0.3.15",
         "helmet": "^7.1.0",
         "https-proxy-agent": "^7.0.2",
         "inquirer": "^8.2.6",
         "ioredis": "^5.3.2",
         "ldapjs": "^3.0.7",
         "morgan": "^1.10.0",
+        "node-cron": "^4.2.1",
         "nodemailer": "^7.0.6",
         "ora": "^5.4.1",
         "rate-limiter-flexible": "^5.0.5",
@@ -44,6 +46,7 @@
         "jest": "^29.7.0",
         "nodemon": "^3.0.1",
         "prettier": "^3.6.2",
+        "prettier-plugin-tailwindcss": "^0.7.2",
         "supertest": "^6.3.3"
       },
       "engines": {
@@ -5396,6 +5399,19 @@
         "node": ">= 0.4"
       }
     },
+    "node_modules/heapdump": {
+      "version": "0.3.15",
+      "resolved": "https://registry.npmjs.org/heapdump/-/heapdump-0.3.15.tgz",
+      "integrity": "sha512-n8aSFscI9r3gfhOcAECAtXFaQ1uy4QSke6bnaL+iymYZ/dWs9cqDqHM+rALfsHUwukUbxsdlECZ0pKmJdQ/4OA==",
+      "hasInstallScript": true,
+      "license": "ISC",
+      "dependencies": {
+        "nan": "^2.13.2"
+      },
+      "engines": {
+        "node": ">=0.10.0"
+      }
+    },
     "node_modules/helmet": {
       "version": "7.2.0",
       "resolved": "https://registry.npmmirror.com/helmet/-/helmet-7.2.0.tgz",
@@ -7011,6 +7027,12 @@
       "integrity": "sha512-nnbWWOkoWyUsTjKrhgD0dcz22mdkSnpYqbEjIm2nhwhuxlSkpywJmBo8h0ZqJdkp73mb90SssHkN4rsRaBAfAA==",
       "license": "ISC"
     },
+    "node_modules/nan": {
+      "version": "2.24.0",
+      "resolved": "https://registry.npmjs.org/nan/-/nan-2.24.0.tgz",
+      "integrity": "sha512-Vpf9qnVW1RaDkoNKFUvfxqAbtI8ncb8OJlqZ9wwpXzWPEsvsB1nvdUi6oYrHIkQ1Y/tMDnr1h4nczS0VB9Xykg==",
+      "license": "MIT"
+    },
     "node_modules/natural-compare": {
       "version": "1.4.0",
       "resolved": "https://registry.npmmirror.com/natural-compare/-/natural-compare-1.4.0.tgz",
@@ -7027,6 +7049,15 @@
         "node": ">= 0.6"
       }
     },
+    "node_modules/node-cron": {
+      "version": "4.2.1",
+      "resolved": "https://registry.npmmirror.com/node-cron/-/node-cron-4.2.1.tgz",
+      "integrity": "sha512-lgimEHPE/QDgFlywTd8yTR61ptugX3Qer29efeyWw2rv259HtGBNn1vZVmp8lB9uo9wC0t/AT4iGqXxia+CJFg==",
+      "license": "ISC",
+      "engines": {
+        "node": ">=6.0.0"
+      }
+    },
     "node_modules/node-domexception": {
       "version": "1.0.0",
       "resolved": "https://registry.npmmirror.com/node-domexception/-/node-domexception-1.0.0.tgz",
@@ -7598,6 +7629,85 @@
         "node": ">=6.0.0"
       }
     },
+    "node_modules/prettier-plugin-tailwindcss": {
+      "version": "0.7.2",
+      "resolved": "https://registry.npmjs.org/prettier-plugin-tailwindcss/-/prettier-plugin-tailwindcss-0.7.2.tgz",
+      "integrity": "sha512-LkphyK3Fw+q2HdMOoiEHWf93fNtYJwfamoKPl7UwtjFQdei/iIBoX11G6j706FzN3ymX9mPVi97qIY8328vdnA==",
+      "dev": true,
+      "license": "MIT",
+      "engines": {
+        "node": ">=20.19"
+      },
+      "peerDependencies": {
+        "@ianvs/prettier-plugin-sort-imports": "*",
+        "@prettier/plugin-hermes": "*",
+        "@prettier/plugin-oxc": "*",
+        "@prettier/plugin-pug": "*",
+        "@shopify/prettier-plugin-liquid": "*",
+        "@trivago/prettier-plugin-sort-imports": "*",
+        "@zackad/prettier-plugin-twig": "*",
+        "prettier": "^3.0",
+        "prettier-plugin-astro": "*",
+        "prettier-plugin-css-order": "*",
+        "prettier-plugin-jsdoc": "*",
+        "prettier-plugin-marko": "*",
+        "prettier-plugin-multiline-arrays": "*",
+        "prettier-plugin-organize-attributes": "*",
+        "prettier-plugin-organize-imports": "*",
+        "prettier-plugin-sort-imports": "*",
+        "prettier-plugin-svelte": "*"
+      },
+      "peerDependenciesMeta": {
+        "@ianvs/prettier-plugin-sort-imports": {
+          "optional": true
+        },
+        "@prettier/plugin-hermes": {
+          "optional": true
+        },
+        "@prettier/plugin-oxc": {
+          "optional": true
+        },
+        "@prettier/plugin-pug": {
+          "optional": true
+        },
+        "@shopify/prettier-plugin-liquid": {
+          "optional": true
+        },
+        "@trivago/prettier-plugin-sort-imports": {
+          "optional": true
+        },
+        "@zackad/prettier-plugin-twig": {
+          "optional": true
+        },
+        "prettier-plugin-astro": {
+          "optional": true
+        },
+        "prettier-plugin-css-order": {
+          "optional": true
+        },
+        "prettier-plugin-jsdoc": {
+          "optional": true
+        },
+        "prettier-plugin-marko": {
+          "optional": true
+        },
+        "prettier-plugin-multiline-arrays": {
+          "optional": true
+        },
+        "prettier-plugin-organize-attributes": {
+          "optional": true
+        },
+        "prettier-plugin-organize-imports": {
+          "optional": true
+        },
+        "prettier-plugin-sort-imports": {
+          "optional": true
+        },
+        "prettier-plugin-svelte": {
+          "optional": true
+        }
+      }
+    },
     "node_modules/pretty-format": {
       "version": "29.7.0",
       "resolved": "https://registry.npmmirror.com/pretty-format/-/pretty-format-29.7.0.tgz",

package.json

@@ -65,6 +65,7 @@
     "ioredis": "^5.3.2",
     "ldapjs": "^3.0.7",
     "morgan": "^1.10.0",
+    "node-cron": "^4.2.1",
     "nodemailer": "^7.0.6",
     "ora": "^5.4.1",
     "rate-limiter-flexible": "^5.0.5",
@@ -83,6 +84,7 @@
     "jest": "^29.7.0",
     "nodemon": "^3.0.1",
     "prettier": "^3.6.2",
+    "prettier-plugin-tailwindcss": "^0.7.2",
     "supertest": "^6.3.3"
   },
   "engines": {

scripts/test-official-models.js

@@ -0,0 +1,108 @@
+#!/usr/bin/env node
+/**
+ * 官方模型版本识别测试 - 最终版 v2
+ */
+
+const { isOpus45OrNewer } = require('../src/utils/modelHelper')
+
+// 官方模型
+const officialModels = [
+  { name: 'claude-3-opus-20240229', desc: 'Opus 3 (已弃用)', expectPro: false },
+  { name: 'claude-opus-4-20250514', desc: 'Opus 4.0', expectPro: false },
+  { name: 'claude-opus-4-1-20250805', desc: 'Opus 4.1', expectPro: false },
+  { name: 'claude-opus-4-5-20251101', desc: 'Opus 4.5', expectPro: true }
+]
+
+// 非 Opus 模型
+const nonOpusModels = [
+  { name: 'claude-sonnet-4-20250514', desc: 'Sonnet 4' },
+  { name: 'claude-sonnet-4-5-20250929', desc: 'Sonnet 4.5' },
+  { name: 'claude-haiku-4-5-20251001', desc: 'Haiku 4.5' },
+  { name: 'claude-3-5-haiku-20241022', desc: 'Haiku 3.5' },
+  { name: 'claude-3-haiku-20240307', desc: 'Haiku 3' },
+  { name: 'claude-3-7-sonnet-20250219', desc: 'Sonnet 3.7 (已弃用)' }
+]
+
+// 其他格式测试
+const otherFormats = [
+  { name: 'claude-opus-4.5', expected: true, desc: 'Opus 4.5 点分隔' },
+  { name: 'claude-opus-4-5', expected: true, desc: 'Opus 4.5 横线分隔' },
+  { name: 'opus-4.5', expected: true, desc: 'Opus 4.5 无前缀' },
+  { name: 'opus-4-5', expected: true, desc: 'Opus 4-5 无前缀' },
+  { name: 'opus-latest', expected: true, desc: 'Opus latest' },
+  { name: 'claude-opus-5', expected: true, desc: 'Opus 5 (未来)' },
+  { name: 'claude-opus-5-0', expected: true, desc: 'Opus 5.0 (未来)' },
+  { name: 'opus-4.0', expected: false, desc: 'Opus 4.0' },
+  { name: 'opus-4.1', expected: false, desc: 'Opus 4.1' },
+  { name: 'opus-4.4', expected: false, desc: 'Opus 4.4' },
+  { name: 'opus-4', expected: false, desc: 'Opus 4' },
+  { name: 'opus-4-0', expected: false, desc: 'Opus 4-0' },
+  { name: 'opus-4-1', expected: false, desc: 'Opus 4-1' },
+  { name: 'opus-4-4', expected: false, desc: 'Opus 4-4' },
+  { name: 'opus', expected: false, desc: '仅 opus' },
+  { name: null, expected: false, desc: 'null' },
+  { name: '', expected: false, desc: '空字符串' }
+]
+
+console.log('='.repeat(90))
+console.log('官方模型版本识别测试 - 最终版 v2')
+console.log('='.repeat(90))
+console.log()
+
+let passed = 0
+let failed = 0
+
+// 测试官方 Opus 模型
+console.log('📌 官方 Opus 模型:')
+for (const m of officialModels) {
+  const result = isOpus45OrNewer(m.name)
+  const status = result === m.expectPro ? '✅ PASS' : '❌ FAIL'
+  if (result === m.expectPro) {
+    passed++
+  } else {
+    failed++
+  }
+  const proSupport = result ? 'Pro 可用 ✅' : 'Pro 不可用 ❌'
+  console.log(`  ${status} | ${m.name.padEnd(32)} | ${m.desc.padEnd(18)} | ${proSupport}`)
+}
+
+console.log()
+console.log('📌 非 Opus 模型 (不受此函数影响):')
+for (const m of nonOpusModels) {
+  const result = isOpus45OrNewer(m.name)
+  console.log(
+    `  ➖      | ${m.name.padEnd(32)} | ${m.desc.padEnd(18)} | ${result ? '⚠️ 异常' : '正确跳过'}`
+  )
+  if (result) {
+    failed++ // 非 Opus 模型不应返回 true
+  }
+}
+
+console.log()
+console.log('📌 其他格式测试:')
+for (const m of otherFormats) {
+  const result = isOpus45OrNewer(m.name)
+  const status = result === m.expected ? '✅ PASS' : '❌ FAIL'
+  if (result === m.expected) {
+    passed++
+  } else {
+    failed++
+  }
+  const display = m.name === null ? 'null' : m.name === '' ? '""' : m.name
+  console.log(
+    `  ${status} | ${display.padEnd(25)} | ${m.desc.padEnd(18)} | ${result ? 'Pro 可用' : 'Pro 不可用'}`
+  )
+}
+
+console.log()
+console.log('='.repeat(90))
+console.log('测试结果:', passed, '通过,', failed, '失败')
+console.log('='.repeat(90))
+
+if (failed > 0) {
+  console.log('\n❌ 有测试失败，请检查函数逻辑')
+  process.exit(1)
+} else {
+  console.log('\n✅ 所有测试通过！函数可以安全使用')
+  process.exit(0)
+}

src/app.js

@@ -52,6 +52,16 @@ class Application {
       await redis.connect()
       logger.success('✅ Redis connected successfully')
 
+      // 💳 初始化账户余额查询服务（Provider 注册）
+      try {
+        const accountBalanceService = require('./services/accountBalanceService')
+        const { registerAllProviders } = require('./services/balanceProviders')
+        registerAllProviders(accountBalanceService)
+        logger.info('✅ 账户余额查询服务已初始化')
+      } catch (error) {
+        logger.warn('⚠️ 账户余额查询服务初始化失败:', error.message)
+      }
+
       // 💰 初始化价格服务
       logger.info('🔄 Initializing pricing service...')
       await pricingService.initialize()
@@ -68,6 +78,10 @@ class Application {
       logger.info('🔄 Initializing admin credentials...')
       await this.initializeAdmin()
 
+      // 🔒 安全启动：清理无效/伪造的管理员会话
+      logger.info('🔒 Cleaning up invalid admin sessions...')
+      await this.cleanupInvalidSessions()
+
       // 💰 初始化费用数据
       logger.info('💰 Checking cost data initialization...')
       const costInitService = require('./services/costInitService')
@@ -165,7 +179,7 @@ class Application {
       // 🔧 基础中间件
       this.app.use(
         express.json({
-          limit: '10mb',
+          limit: '100mb',
           verify: (req, res, buf, encoding) => {
             // 验证JSON格式
             if (buf && buf.length && !buf.toString(encoding || 'utf8').trim()) {
@@ -174,7 +188,7 @@ class Application {
           }
         })
       )
-      this.app.use(express.urlencoded({ extended: true, limit: '10mb' }))
+      this.app.use(express.urlencoded({ extended: true, limit: '100mb' }))
       this.app.use(securityMiddleware)
 
       // 🎯 信任代理
@@ -264,6 +278,25 @@ class Application {
       this.app.use('/api', apiRoutes)
       this.app.use('/api', unifiedRoutes) // 统一智能路由（支持 /v1/chat/completions 等）
       this.app.use('/claude', apiRoutes) // /claude 路由别名，与 /api 功能相同
+      // Anthropic (Claude Code) 路由：按路径强制分流到 Gemini OAuth 账户
+      // - /antigravity/api/v1/messages -> Antigravity OAuth
+      // - /gemini-cli/api/v1/messages -> Gemini CLI OAuth
+      this.app.use(
+        '/antigravity/api',
+        (req, res, next) => {
+          req._anthropicVendor = 'antigravity'
+          next()
+        },
+        apiRoutes
+      )
+      this.app.use(
+        '/gemini-cli/api',
+        (req, res, next) => {
+          req._anthropicVendor = 'gemini-cli'
+          next()
+        },
+        apiRoutes
+      )
       this.app.use('/admin', adminRoutes)
       this.app.use('/users', userRoutes)
       // 使用 web 路由（包含 auth 和页面重定向）
@@ -426,6 +459,54 @@ class Application {
     }
   }
 
+  // 🔒 清理无效/伪造的管理员会话（安全启动检查）
+  async cleanupInvalidSessions() {
+    try {
+      const client = redis.getClient()
+
+      // 获取所有 session:* 键
+      const sessionKeys = await client.keys('session:*')
+
+      let validCount = 0
+      let invalidCount = 0
+
+      for (const key of sessionKeys) {
+        // 跳过 admin_credentials（系统凭据）
+        if (key === 'session:admin_credentials') {
+          continue
+        }
+
+        const sessionData = await client.hgetall(key)
+
+        // 检查会话完整性：必须有 username 和 loginTime
+        const hasUsername = !!sessionData.username
+        const hasLoginTime = !!sessionData.loginTime
+
+        if (!hasUsername || !hasLoginTime) {
+          // 无效会话 - 可能是漏洞利用创建的伪造会话
+          invalidCount++
+          logger.security(
+            `🔒 Removing invalid session: ${key} (username: ${hasUsername}, loginTime: ${hasLoginTime})`
+          )
+          await client.del(key)
+        } else {
+          validCount++
+        }
+      }
+
+      if (invalidCount > 0) {
+        logger.security(`🔒 Startup security check: Removed ${invalidCount} invalid sessions`)
+      }
+
+      logger.success(
+        `✅ Session cleanup completed: ${validCount} valid, ${invalidCount} invalid removed`
+      )
+    } catch (error) {
+      // 清理失败不应阻止服务启动
+      logger.error('❌ Failed to cleanup invalid sessions:', error.message)
+    }
+  }
+
   // 🔍 Redis健康检查
   async checkRedisHealth() {
     try {
@@ -581,15 +662,40 @@ class Application {
 
         const now = Date.now()
         let totalCleaned = 0
+        let legacyCleaned = 0
 
         // 使用 Lua 脚本批量清理所有过期项
         for (const key of keys) {
+          // 跳过已知非 Sorted Set 类型的键（这些键有各自的清理逻辑）
+          // - concurrency:queue:stats:* 是 Hash 类型
+          // - concurrency:queue:wait_times:* 是 List 类型
+          // - concurrency:queue:* (不含stats/wait_times) 是 String 类型
+          if (
+            key.startsWith('concurrency:queue:stats:') ||
+            key.startsWith('concurrency:queue:wait_times:') ||
+            (key.startsWith('concurrency:queue:') &&
+              !key.includes(':stats:') &&
+              !key.includes(':wait_times:'))
+          ) {
+            continue
+          }
+
           try {
-            const cleaned = await redis.client.eval(
+            // 使用原子 Lua 脚本：先检查类型，再执行清理
+            // 返回值：0 = 正常清理无删除，1 = 清理后删除空键，-1 = 遗留键已删除
+            const result = await redis.client.eval(
               `
               local key = KEYS[1]
               local now = tonumber(ARGV[1])
 
+              -- 先检查键类型，只对 Sorted Set 执行清理
+              local keyType = redis.call('TYPE', key)
+              if keyType.ok ~= 'zset' then
+                -- 非 ZSET 类型的遗留键，直接删除
+                redis.call('DEL', key)
+                return -1
+              end
+
               -- 清理过期项
               redis.call('ZREMRANGEBYSCORE', key, '-inf', now)
 
@@ -608,8 +714,10 @@ class Application {
               key,
               now
             )
-            if (cleaned === 1) {
+            if (result === 1) {
               totalCleaned++
+            } else if (result === -1) {
+              legacyCleaned++
             }
           } catch (error) {
             logger.error(`❌ Failed to clean concurrency key ${key}:`, error)
@@ -619,12 +727,50 @@ class Application {
         if (totalCleaned > 0) {
           logger.info(`🔢 Concurrency cleanup: cleaned ${totalCleaned} expired keys`)
         }
+        if (legacyCleaned > 0) {
+          logger.warn(`🧹 Concurrency cleanup: removed ${legacyCleaned} legacy keys (wrong type)`)
+        }
       } catch (error) {
         logger.error('❌ Concurrency cleanup task failed:', error)
       }
     }, 60000) // 每分钟执行一次
 
     logger.info('🔢 Concurrency cleanup task started (running every 1 minute)')
+
+    // 📬 启动用户消息队列服务
+    const userMessageQueueService = require('./services/userMessageQueueService')
+    // 先清理服务重启后残留的锁，防止旧锁阻塞新请求
+    userMessageQueueService.cleanupStaleLocks().then(() => {
+      // 然后启动定时清理任务
+      userMessageQueueService.startCleanupTask()
+    })
+
+    // 🚦 清理服务重启后残留的并发排队计数器
+    // 多实例部署时建议关闭此开关，避免新实例启动时清空其他实例的队列计数
+    // 可通过 DELETE /admin/concurrency/queue 接口手动清理
+    const clearQueuesOnStartup = process.env.CLEAR_CONCURRENCY_QUEUES_ON_STARTUP !== 'false'
+    if (clearQueuesOnStartup) {
+      redis.clearAllConcurrencyQueues().catch((error) => {
+        logger.error('❌ Error clearing concurrency queues on startup:', error)
+      })
+    } else {
+      logger.info(
+        '🚦 Skipping concurrency queue cleanup on startup (CLEAR_CONCURRENCY_QUEUES_ON_STARTUP=false)'
+      )
+    }
+
+    // 🧪 启动账户定时测试调度器
+    // 根据配置定期测试账户连通性并保存测试历史
+    const accountTestSchedulerEnabled =
+      process.env.ACCOUNT_TEST_SCHEDULER_ENABLED !== 'false' &&
+      config.accountTestScheduler?.enabled !== false
+    if (accountTestSchedulerEnabled) {
+      const accountTestSchedulerService = require('./services/accountTestSchedulerService')
+      accountTestSchedulerService.start()
+      logger.info('🧪 Account test scheduler service started')
+    } else {
+      logger.info('🧪 Account test scheduler service disabled')
+    }
   }
 
   setupGracefulShutdown() {
@@ -661,6 +807,15 @@ class Application {
             logger.error('❌ Error stopping rate limit cleanup service:', error)
           }
 
+          // 停止用户消息队列清理服务
+          try {
+            const userMessageQueueService = require('./services/userMessageQueueService')
+            userMessageQueueService.stopCleanupTask()
+            logger.info('📬 User message queue service stopped')
+          } catch (error) {
+            logger.error('❌ Error stopping user message queue service:', error)
+          }
+
           // 停止费用排序索引服务
           try {
             const costRankService = require('./services/costRankService')
@@ -670,6 +825,15 @@ class Application {
             logger.error('❌ Error stopping cost rank service:', error)
           }
 
+          // 停止账户定时测试调度器
+          try {
+            const accountTestSchedulerService = require('./services/accountTestSchedulerService')
+            accountTestSchedulerService.stop()
+            logger.info('🧪 Account test scheduler service stopped')
+          } catch (error) {
+            logger.error('❌ Error stopping account test scheduler service:', error)
+          }
+
           // 🔢 清理所有并发计数（Phase 1 修复：防止重启泄漏）
           try {
             logger.info('🔢 Cleaning up all concurrency counters...')

src/handlers/geminiHandlers.js

@@ -9,6 +9,7 @@ const logger = require('../utils/logger')
 const geminiAccountService = require('../services/geminiAccountService')
 const geminiApiAccountService = require('../services/geminiApiAccountService')
 const { sendGeminiRequest, getAvailableModels } = require('../services/geminiRelayService')
+const { sendAntigravityRequest } = require('../services/antigravityRelayService')
 const crypto = require('crypto')
 const sessionHelper = require('../utils/sessionHelper')
 const unifiedGeminiScheduler = require('../services/unifiedGeminiScheduler')
@@ -86,8 +87,7 @@ function generateSessionHash(req) {
  * 检查 API Key 权限
  */
 function checkPermissions(apiKeyData, requiredPermission = 'gemini') {
-  const permissions = apiKeyData?.permissions || 'all'
-  return permissions === 'all' || permissions === requiredPermission
+  return apiKeyService.hasPermission(apiKeyData?.permissions, requiredPermission)
 }
 
 /**
@@ -449,9 +449,8 @@ async function handleMessages(req, res) {
 
       // 添加代理配置
       if (proxyConfig) {
-        const proxyHelper = new ProxyHelper()
-        axiosConfig.httpsAgent = proxyHelper.createProxyAgent(proxyConfig)
-        axiosConfig.httpAgent = proxyHelper.createProxyAgent(proxyConfig)
+        axiosConfig.httpsAgent = ProxyHelper.createProxyAgent(proxyConfig)
+        axiosConfig.httpAgent = ProxyHelper.createProxyAgent(proxyConfig)
       }
 
       try {
@@ -509,20 +508,37 @@ async function handleMessages(req, res) {
       // OAuth 账户：使用现有的 sendGeminiRequest
       // 智能处理项目ID：优先使用配置的 projectId，降级到临时 tempProjectId
       const effectiveProjectId = account.projectId || account.tempProjectId || null
+      const oauthProvider = account.oauthProvider || 'gemini-cli'
 
-      geminiResponse = await sendGeminiRequest({
-        messages,
-        model,
-        temperature,
-        maxTokens: max_tokens,
-        stream,
-        accessToken: account.accessToken,
-        proxy: account.proxy,
-        apiKeyId: apiKeyData.id,
-        signal: abortController.signal,
-        projectId: effectiveProjectId,
-        accountId: account.id
-      })
+      if (oauthProvider === 'antigravity') {
+        geminiResponse = await sendAntigravityRequest({
+          messages,
+          model,
+          temperature,
+          maxTokens: max_tokens,
+          stream,
+          accessToken: account.accessToken,
+          proxy: account.proxy,
+          apiKeyId: apiKeyData.id,
+          signal: abortController.signal,
+          projectId: effectiveProjectId,
+          accountId: account.id
+        })
+      } else {
+        geminiResponse = await sendGeminiRequest({
+          messages,
+          model,
+          temperature,
+          maxTokens: max_tokens,
+          stream,
+          accessToken: account.accessToken,
+          proxy: account.proxy,
+          apiKeyId: apiKeyData.id,
+          signal: abortController.signal,
+          projectId: effectiveProjectId,
+          accountId: account.id
+        })
+      }
     }
 
     if (stream) {
@@ -732,9 +748,8 @@ async function handleModels(req, res) {
           headers: { 'Content-Type': 'application/json' }
         }
         if (proxyConfig) {
-          const proxyHelper = new ProxyHelper()
-          axiosConfig.httpsAgent = proxyHelper.createProxyAgent(proxyConfig)
-          axiosConfig.httpAgent = proxyHelper.createProxyAgent(proxyConfig)
+          axiosConfig.httpsAgent = ProxyHelper.createProxyAgent(proxyConfig)
+          axiosConfig.httpAgent = ProxyHelper.createProxyAgent(proxyConfig)
         }
         const response = await axios(axiosConfig)
         models = (response.data.models || []).map((m) => ({
@@ -756,8 +771,16 @@ async function handleModels(req, res) {
         ]
       }
     } else {
-      // OAuth 账户：使用 OAuth token 获取模型列表
-      models = await getAvailableModels(account.accessToken, account.proxy)
+      // OAuth 账户：根据 OAuth provider 选择上游
+      const oauthProvider = account.oauthProvider || 'gemini-cli'
+      models =
+        oauthProvider === 'antigravity'
+          ? await geminiAccountService.fetchAvailableModelsAntigravity(
+              account.accessToken,
+              account.proxy,
+              account.refreshToken
+            )
+          : await getAvailableModels(account.accessToken, account.proxy)
     }
 
     res.json({
@@ -839,7 +862,7 @@ async function handleKeyInfo(req, res) {
     res.json({
       id: keyData.id,
       name: keyData.name,
-      permissions: keyData.permissions || 'all',
+      permissions: keyData.permissions,
       token_limit: keyData.tokenLimit,
       tokens_used: keyData.usage.total.tokens,
       tokens_remaining:
@@ -929,7 +952,8 @@ function handleSimpleEndpoint(apiMethod) {
       const client = await geminiAccountService.getOauthClient(
         accessToken,
         refreshToken,
-        proxyConfig
+        proxyConfig,
+        account.oauthProvider
       )
 
       // 直接转发请求体，不做特殊处理
@@ -1008,7 +1032,12 @@ async function handleLoadCodeAssist(req, res) {
     // 解析账户的代理配置
     const proxyConfig = parseProxyConfig(account)
 
-    const client = await geminiAccountService.getOauthClient(accessToken, refreshToken, proxyConfig)
+    const client = await geminiAccountService.getOauthClient(
+      accessToken,
+      refreshToken,
+      proxyConfig,
+      account.oauthProvider
+    )
 
     // 智能处理项目ID
     const effectiveProjectId = projectId || cloudaicompanionProject || null
@@ -1106,7 +1135,12 @@ async function handleOnboardUser(req, res) {
     // 解析账户的代理配置
     const proxyConfig = parseProxyConfig(account)
 
-    const client = await geminiAccountService.getOauthClient(accessToken, refreshToken, proxyConfig)
+    const client = await geminiAccountService.getOauthClient(
+      accessToken,
+      refreshToken,
+      proxyConfig,
+      account.oauthProvider
+    )
 
     // 智能处理项目ID
     const effectiveProjectId = projectId || cloudaicompanionProject || null
@@ -1154,6 +1188,110 @@ async function handleOnboardUser(req, res) {
   }
 }
 
+/**
+ * 处理 retrieveUserQuota 请求
+ * POST /v1internal:retrieveUserQuota
+ *
+ * 功能：查询用户在各个Gemini模型上的配额使用情况
+ * 请求体：{ "project": "项目ID" }
+ * 响应：{ "buckets": [...] }
+ */
+async function handleRetrieveUserQuota(req, res) {
+  try {
+    // 1. 权限检查
+    if (!ensureGeminiPermission(req, res)) {
+      return undefined
+    }
+
+    // 2. 会话哈希
+    const sessionHash = sessionHelper.generateSessionHash(req.body)
+
+    // 3. 账户选择
+    const requestedModel = req.body.model || req.params.modelName || 'gemini-2.5-flash'
+    const schedulerResult = await unifiedGeminiScheduler.selectAccountForApiKey(
+      req.apiKey,
+      sessionHash,
+      requestedModel
+    )
+    const { accountId, accountType } = schedulerResult
+
+    // 4. 账户类型验证 - v1internal 路由只支持 OAuth 账户
+    if (accountType === 'gemini-api') {
+      logger.error(`❌ v1internal routes do not support Gemini API accounts. Account: ${accountId}`)
+      return res.status(400).json({
+        error: {
+          message:
+            'This endpoint only supports Gemini OAuth accounts. Gemini API Key accounts are not compatible with v1internal format.',
+          type: 'invalid_account_type'
+        }
+      })
+    }
+
+    // 5. 获取账户
+    const account = await geminiAccountService.getAccount(accountId)
+    if (!account) {
+      return res.status(404).json({
+        error: {
+          message: 'Gemini account not found',
+          type: 'account_not_found'
+        }
+      })
+    }
+    const { accessToken, refreshToken, projectId } = account
+
+    // 6. 从请求体提取项目字段（注意：字段名是 "project"，不是 "cloudaicompanionProject"）
+    const requestProject = req.body.project
+
+    const version = req.path.includes('v1beta') ? 'v1beta' : 'v1internal'
+    logger.info(`RetrieveUserQuota request (${version})`, {
+      requestedProject: requestProject || null,
+      accountProject: projectId || null,
+      apiKeyId: req.apiKey?.id || 'unknown'
+    })
+
+    // 7. 解析账户的代理配置
+    const proxyConfig = parseProxyConfig(account)
+
+    // 8. 获取OAuth客户端
+    const client = await geminiAccountService.getOauthClient(accessToken, refreshToken, proxyConfig)
+
+    // 9. 智能处理项目ID（与其他 v1internal 接口保持一致）
+    const effectiveProject = projectId || requestProject || null
+
+    logger.info('📋 retrieveUserQuota项目ID处理逻辑', {
+      accountProjectId: projectId,
+      requestProject,
+      effectiveProject,
+      decision: projectId ? '使用账户配置' : requestProject ? '使用请求参数' : '不使用项目ID'
+    })
+
+    // 10. 构建请求体（注入 effectiveProject）
+    const requestBody = { ...req.body }
+    if (effectiveProject) {
+      requestBody.project = effectiveProject
+    }
+
+    // 11. 调用底层服务转发请求
+    const response = await geminiAccountService.forwardToCodeAssist(
+      client,
+      'retrieveUserQuota',
+      requestBody,
+      proxyConfig
+    )
+
+    res.json(response)
+  } catch (error) {
+    const version = req.path.includes('v1beta') ? 'v1beta' : 'v1internal'
+    logger.error(`Error in retrieveUserQuota endpoint (${version})`, {
+      error: error.message
+    })
+    res.status(500).json({
+      error: 'Internal server error',
+      message: error.message
+    })
+  }
+}
+
 /**
  * 处理 countTokens 请求
  */
@@ -1234,9 +1372,8 @@ async function handleCountTokens(req, res) {
       }
 
       if (proxyConfig) {
-        const proxyHelper = new ProxyHelper()
-        axiosConfig.httpsAgent = proxyHelper.createProxyAgent(proxyConfig)
-        axiosConfig.httpAgent = proxyHelper.createProxyAgent(proxyConfig)
+        axiosConfig.httpsAgent = ProxyHelper.createProxyAgent(proxyConfig)
+        axiosConfig.httpAgent = ProxyHelper.createProxyAgent(proxyConfig)
       }
 
       try {
@@ -1259,7 +1396,8 @@ async function handleCountTokens(req, res) {
       const client = await geminiAccountService.getOauthClient(
         accessToken,
         refreshToken,
-        proxyConfig
+        proxyConfig,
+        account.oauthProvider
       )
       response = await geminiAccountService.countTokens(client, contents, model, proxyConfig)
     }
@@ -1369,13 +1507,20 @@ async function handleGenerateContent(req, res) {
     // 解析账户的代理配置
     const proxyConfig = parseProxyConfig(account)
 
-    const client = await geminiAccountService.getOauthClient(accessToken, refreshToken, proxyConfig)
+    const client = await geminiAccountService.getOauthClient(
+      accessToken,
+      refreshToken,
+      proxyConfig,
+      account.oauthProvider
+    )
 
     // 智能处理项目ID：优先使用配置的 projectId，降级到临时 tempProjectId
     let effectiveProjectId = account.projectId || account.tempProjectId || null
 
+    const oauthProvider = account.oauthProvider || 'gemini-cli'
+
     // 如果没有任何项目ID，尝试调用 loadCodeAssist 获取
-    if (!effectiveProjectId) {
+    if (!effectiveProjectId && oauthProvider !== 'antigravity') {
       try {
         logger.info('📋 No projectId available, attempting to fetch from loadCodeAssist...')
         const loadResponse = await geminiAccountService.loadCodeAssist(client, null, proxyConfig)
@@ -1391,6 +1536,12 @@ async function handleGenerateContent(req, res) {
       }
     }
 
+    if (!effectiveProjectId && oauthProvider === 'antigravity') {
+      // Antigravity 账号允许没有 projectId：生成一个稳定的临时 projectId 并缓存
+      effectiveProjectId = `ag-${crypto.randomUUID().replace(/-/g, '').slice(0, 16)}`
+      await geminiAccountService.updateTempProjectId(accountId, effectiveProjectId)
+    }
+
     // 如果还是没有项目ID，返回错误
     if (!effectiveProjectId) {
       return res.status(403).json({
@@ -1413,14 +1564,24 @@ async function handleGenerateContent(req, res) {
           : '从loadCodeAssist获取'
     })
 
-    const response = await geminiAccountService.generateContent(
-      client,
-      { model, request: actualRequestData },
-      user_prompt_id,
-      effectiveProjectId,
-      req.apiKey?.id,
-      proxyConfig
-    )
+    const response =
+      oauthProvider === 'antigravity'
+        ? await geminiAccountService.generateContentAntigravity(
+            client,
+            { model, request: actualRequestData },
+            user_prompt_id,
+            effectiveProjectId,
+            req.apiKey?.id,
+            proxyConfig
+          )
+        : await geminiAccountService.generateContent(
+            client,
+            { model, request: actualRequestData },
+            user_prompt_id,
+            effectiveProjectId,
+            req.apiKey?.id,
+            proxyConfig
+          )
 
     // 记录使用统计
     if (response?.response?.usageMetadata) {
@@ -1581,13 +1742,20 @@ async function handleStreamGenerateContent(req, res) {
     // 解析账户的代理配置
     const proxyConfig = parseProxyConfig(account)
 
-    const client = await geminiAccountService.getOauthClient(accessToken, refreshToken, proxyConfig)
+    const client = await geminiAccountService.getOauthClient(
+      accessToken,
+      refreshToken,
+      proxyConfig,
+      account.oauthProvider
+    )
 
     // 智能处理项目ID：优先使用配置的 projectId，降级到临时 tempProjectId
     let effectiveProjectId = account.projectId || account.tempProjectId || null
 
+    const oauthProvider = account.oauthProvider || 'gemini-cli'
+
     // 如果没有任何项目ID，尝试调用 loadCodeAssist 获取
-    if (!effectiveProjectId) {
+    if (!effectiveProjectId && oauthProvider !== 'antigravity') {
       try {
         logger.info('📋 No projectId available, attempting to fetch from loadCodeAssist...')
         const loadResponse = await geminiAccountService.loadCodeAssist(client, null, proxyConfig)
@@ -1603,6 +1771,11 @@ async function handleStreamGenerateContent(req, res) {
       }
     }
 
+    if (!effectiveProjectId && oauthProvider === 'antigravity') {
+      effectiveProjectId = `ag-${crypto.randomUUID().replace(/-/g, '').slice(0, 16)}`
+      await geminiAccountService.updateTempProjectId(accountId, effectiveProjectId)
+    }
+
     // 如果还是没有项目ID，返回错误
     if (!effectiveProjectId) {
       return res.status(403).json({
@@ -1625,15 +1798,26 @@ async function handleStreamGenerateContent(req, res) {
           : '从loadCodeAssist获取'
     })
 
-    const streamResponse = await geminiAccountService.generateContentStream(
-      client,
-      { model, request: actualRequestData },
-      user_prompt_id,
-      effectiveProjectId,
-      req.apiKey?.id,
-      abortController.signal,
-      proxyConfig
-    )
+    const streamResponse =
+      oauthProvider === 'antigravity'
+        ? await geminiAccountService.generateContentStreamAntigravity(
+            client,
+            { model, request: actualRequestData },
+            user_prompt_id,
+            effectiveProjectId,
+            req.apiKey?.id,
+            abortController.signal,
+            proxyConfig
+          )
+        : await geminiAccountService.generateContentStream(
+            client,
+            { model, request: actualRequestData },
+            user_prompt_id,
+            effectiveProjectId,
+            req.apiKey?.id,
+            abortController.signal,
+            proxyConfig
+          )
 
     // 设置 SSE 响应头
     res.setHeader('Content-Type', 'text/event-stream')
@@ -1963,9 +2147,8 @@ async function handleStandardGenerateContent(req, res) {
       }
 
       if (proxyConfig) {
-        const proxyHelper = new ProxyHelper()
-        axiosConfig.httpsAgent = proxyHelper.createProxyAgent(proxyConfig)
-        axiosConfig.httpAgent = proxyHelper.createProxyAgent(proxyConfig)
+        axiosConfig.httpsAgent = ProxyHelper.createProxyAgent(proxyConfig)
+        axiosConfig.httpAgent = ProxyHelper.createProxyAgent(proxyConfig)
       }
 
       try {
@@ -1982,15 +2165,23 @@ async function handleStandardGenerateContent(req, res) {
     } else {
       // OAuth 账户
       const { accessToken, refreshToken } = account
+      const oauthProvider = account.oauthProvider || 'gemini-cli'
       const client = await geminiAccountService.getOauthClient(
         accessToken,
         refreshToken,
-        proxyConfig
+        proxyConfig,
+        oauthProvider
       )
 
       let effectiveProjectId = account.projectId || account.tempProjectId || null
 
-      if (!effectiveProjectId) {
+      if (oauthProvider === 'antigravity') {
+        if (!effectiveProjectId) {
+          // Antigravity 账号允许没有 projectId：生成一个稳定的临时 projectId 并缓存
+          effectiveProjectId = `ag-${crypto.randomUUID().replace(/-/g, '').slice(0, 16)}`
+          await geminiAccountService.updateTempProjectId(actualAccountId, effectiveProjectId)
+        }
+      } else if (!effectiveProjectId) {
         try {
           logger.info('📋 No projectId available, attempting to fetch from loadCodeAssist...')
           const loadResponse = await geminiAccountService.loadCodeAssist(client, null, proxyConfig)
@@ -2028,14 +2219,25 @@ async function handleStandardGenerateContent(req, res) {
 
       const userPromptId = `${crypto.randomUUID()}########0`
 
-      response = await geminiAccountService.generateContent(
-        client,
-        { model, request: actualRequestData },
-        userPromptId,
-        effectiveProjectId,
-        req.apiKey?.id,
-        proxyConfig
-      )
+      if (oauthProvider === 'antigravity') {
+        response = await geminiAccountService.generateContentAntigravity(
+          client,
+          { model, request: actualRequestData },
+          userPromptId,
+          effectiveProjectId,
+          req.apiKey?.id,
+          proxyConfig
+        )
+      } else {
+        response = await geminiAccountService.generateContent(
+          client,
+          { model, request: actualRequestData },
+          userPromptId,
+          effectiveProjectId,
+          req.apiKey?.id,
+          proxyConfig
+        )
+      }
     }
 
     // 记录使用统计
@@ -2246,9 +2448,8 @@ async function handleStandardStreamGenerateContent(req, res) {
       }
 
       if (proxyConfig) {
-        const proxyHelper = new ProxyHelper()
-        axiosConfig.httpsAgent = proxyHelper.createProxyAgent(proxyConfig)
-        axiosConfig.httpAgent = proxyHelper.createProxyAgent(proxyConfig)
+        axiosConfig.httpsAgent = ProxyHelper.createProxyAgent(proxyConfig)
+        axiosConfig.httpAgent = ProxyHelper.createProxyAgent(proxyConfig)
       }
 
       try {
@@ -2268,12 +2469,20 @@ async function handleStandardStreamGenerateContent(req, res) {
       const client = await geminiAccountService.getOauthClient(
         accessToken,
         refreshToken,
-        proxyConfig
+        proxyConfig,
+        account.oauthProvider
       )
 
       let effectiveProjectId = account.projectId || account.tempProjectId || null
 
-      if (!effectiveProjectId) {
+      const oauthProvider = account.oauthProvider || 'gemini-cli'
+
+      if (oauthProvider === 'antigravity') {
+        if (!effectiveProjectId) {
+          effectiveProjectId = `ag-${crypto.randomUUID().replace(/-/g, '').slice(0, 16)}`
+          await geminiAccountService.updateTempProjectId(actualAccountId, effectiveProjectId)
+        }
+      } else if (!effectiveProjectId) {
         try {
           logger.info('📋 No projectId available, attempting to fetch from loadCodeAssist...')
           const loadResponse = await geminiAccountService.loadCodeAssist(client, null, proxyConfig)
@@ -2311,15 +2520,27 @@ async function handleStandardStreamGenerateContent(req, res) {
 
       const userPromptId = `${crypto.randomUUID()}########0`
 
-      streamResponse = await geminiAccountService.generateContentStream(
-        client,
-        { model, request: actualRequestData },
-        userPromptId,
-        effectiveProjectId,
-        req.apiKey?.id,
-        abortController.signal,
-        proxyConfig
-      )
+      if (oauthProvider === 'antigravity') {
+        streamResponse = await geminiAccountService.generateContentStreamAntigravity(
+          client,
+          { model, request: actualRequestData },
+          userPromptId,
+          effectiveProjectId,
+          req.apiKey?.id,
+          abortController.signal,
+          proxyConfig
+        )
+      } else {
+        streamResponse = await geminiAccountService.generateContentStream(
+          client,
+          { model, request: actualRequestData },
+          userPromptId,
+          effectiveProjectId,
+          req.apiKey?.id,
+          abortController.signal,
+          proxyConfig
+        )
+      }
     }
 
     // 设置 SSE 响应头
@@ -2581,6 +2802,7 @@ module.exports = {
   handleSimpleEndpoint,
   handleLoadCodeAssist,
   handleOnboardUser,
+  handleRetrieveUserQuota,
   handleCountTokens,
   handleGenerateContent,
   handleStreamGenerateContent,

src/middleware/auth.js

@@ -6,6 +6,104 @@ const logger = require('../utils/logger')
 const redis = require('../models/redis')
 // const { RateLimiterRedis } = require('rate-limiter-flexible') // 暂时未使用
 const ClientValidator = require('../validators/clientValidator')
+const ClaudeCodeValidator = require('../validators/clients/claudeCodeValidator')
+const claudeRelayConfigService = require('../services/claudeRelayConfigService')
+const { calculateWaitTimeStats } = require('../utils/statsHelper')
+
+// 工具函数
+function sleep(ms) {
+  return new Promise((resolve) => setTimeout(resolve, ms))
+}
+
+/**
+ * 检查排队是否过载，决定是否应该快速失败
+ * 详见 design.md Decision 7: 排队健康检查与快速失败
+ *
+ * @param {string} apiKeyId - API Key ID
+ * @param {number} timeoutMs - 排队超时时间（毫秒）
+ * @param {Object} queueConfig - 队列配置
+ * @param {number} maxQueueSize - 最大排队数
+ * @returns {Promise<Object>} { reject: boolean, reason?: string, estimatedWaitMs?: number, timeoutMs?: number }
+ */
+async function shouldRejectDueToOverload(apiKeyId, timeoutMs, queueConfig, maxQueueSize) {
+  try {
+    // 如果健康检查被禁用，直接返回不拒绝
+    if (!queueConfig.concurrentRequestQueueHealthCheckEnabled) {
+      return { reject: false, reason: 'health_check_disabled' }
+    }
+
+    // 🔑 先检查当前队列长度
+    const currentQueueCount = await redis.getConcurrencyQueueCount(apiKeyId).catch(() => 0)
+
+    // 队列为空，说明系统已恢复，跳过健康检查
+    if (currentQueueCount === 0) {
+      return { reject: false, reason: 'queue_empty', currentQueueCount: 0 }
+    }
+
+    // 🔑 关键改进：只有当队列接近满载时才进行健康检查
+    // 队列长度 <= maxQueueSize * 0.5 时，认为系统有足够余量，跳过健康检查
+    // 这避免了在队列较短时过于保守地拒绝请求
+    // 使用 ceil 确保小队列（如 maxQueueSize=3）时阈值为 2，即队列 <=1 时跳过
+    const queueLoadThreshold = Math.ceil(maxQueueSize * 0.5)
+    if (currentQueueCount <= queueLoadThreshold) {
+      return {
+        reject: false,
+        reason: 'queue_not_loaded',
+        currentQueueCount,
+        queueLoadThreshold,
+        maxQueueSize
+      }
+    }
+
+    // 获取该 API Key 的等待时间样本
+    const waitTimes = await redis.getQueueWaitTimes(apiKeyId)
+    const stats = calculateWaitTimeStats(waitTimes)
+
+    // 样本不足（< 10），跳过健康检查，避免冷启动误判
+    if (!stats || stats.sampleCount < 10) {
+      return { reject: false, reason: 'insufficient_samples', sampleCount: stats?.sampleCount || 0 }
+    }
+
+    // P90 不可靠时也跳过（虽然 sampleCount >= 10 时 p90Unreliable 应该是 false）
+    if (stats.p90Unreliable) {
+      return { reject: false, reason: 'p90_unreliable', sampleCount: stats.sampleCount }
+    }
+
+    // 计算健康阈值：P90 >= 超时时间 × 阈值 时拒绝
+    const threshold = queueConfig.concurrentRequestQueueHealthThreshold || 0.8
+    const maxAllowedP90 = timeoutMs * threshold
+
+    if (stats.p90 >= maxAllowedP90) {
+      return {
+        reject: true,
+        reason: 'queue_overloaded',
+        estimatedWaitMs: stats.p90,
+        timeoutMs,
+        threshold,
+        sampleCount: stats.sampleCount,
+        currentQueueCount,
+        maxQueueSize
+      }
+    }
+
+    return { reject: false, p90: stats.p90, sampleCount: stats.sampleCount, currentQueueCount }
+  } catch (error) {
+    // 健康检查出错时不阻塞请求，记录警告并继续
+    logger.warn(`Health check failed for ${apiKeyId}:`, error.message)
+    return { reject: false, reason: 'health_check_error', error: error.message }
+  }
+}
+
+// 排队轮询配置常量（可通过配置文件覆盖）
+// 性能权衡：初始间隔越短响应越快，但 Redis QPS 越高
+// 当前配置：100 个等待者时约 250-300 QPS（指数退避后）
+const QUEUE_POLLING_CONFIG = {
+  pollIntervalMs: 200, // 初始轮询间隔（毫秒）- 平衡响应速度和 Redis 压力
+  maxPollIntervalMs: 2000, // 最大轮询间隔（毫秒）- 长时间等待时降低 Redis 压力
+  backoffFactor: 1.5, // 指数退避系数
+  jitterRatio: 0.2, // 抖动比例（±20%）- 防止惊群效应
+  maxRedisFailCount: 5 // 连续 Redis 失败阈值（从 3 提高到 5，提高网络抖动容忍度）
+}
 
 const FALLBACK_CONCURRENCY_CONFIG = {
   leaseSeconds: 300,
@@ -126,9 +224,223 @@ function isTokenCountRequest(req) {
   return false
 }
 
+/**
+ * 等待并发槽位（排队机制核心）
+ *
+ * 采用「先占后检查」模式避免竞态条件：
+ * - 每次轮询时尝试 incrConcurrency 占位
+ * - 如果超限则 decrConcurrency 释放并继续等待
+ * - 成功获取槽位后返回，调用方无需再次 incrConcurrency
+ *
+ * ⚠️ 重要清理责任说明：
+ * - 排队计数：此函数的 finally 块负责调用 decrConcurrencyQueue 清理
+ * - 并发槽位：当返回 acquired=true 时，槽位已被占用（通过 incrConcurrency）
+ *   调用方必须在请求结束时调用 decrConcurrency 释放槽位
+ *   （已在 authenticateApiKey 的 finally 块中处理）
+ *
+ * @param {Object} req - Express 请求对象
+ * @param {Object} res - Express 响应对象
+ * @param {string} apiKeyId - API Key ID
+ * @param {Object} queueOptions - 配置参数
+ * @returns {Promise<Object>} { acquired: boolean, reason?: string, waitTimeMs: number }
+ */
+async function waitForConcurrencySlot(req, res, apiKeyId, queueOptions) {
+  const {
+    concurrencyLimit,
+    requestId,
+    leaseSeconds,
+    timeoutMs,
+    pollIntervalMs,
+    maxPollIntervalMs,
+    backoffFactor,
+    jitterRatio,
+    maxRedisFailCount: configMaxRedisFailCount
+  } = queueOptions
+
+  let clientDisconnected = false
+  // 追踪轮询过程中是否临时占用了槽位（用于异常时清理）
+  // 工作流程：
+  // 1. incrConcurrency 成功且 count <= limit 时，设置 internalSlotAcquired = true
+  // 2. 统计记录完成后，设置 internalSlotAcquired = false 并返回（所有权转移给调用方）
+  // 3. 如果在步骤 1-2 之间发生异常，finally 块会检测到 internalSlotAcquired = true 并释放槽位
+  let internalSlotAcquired = false
+
+  // 监听客户端断开事件
+  // ⚠️ 重要：必须监听 socket 的事件，而不是 req 的事件！
+  // 原因：对于 POST 请求，当 body-parser 读取完请求体后，req（IncomingMessage 可读流）
+  // 的 'close' 事件会立即触发，但这不代表客户端断开连接！客户端仍在等待响应。
+  // socket 的 'close' 事件才是真正的连接关闭信号。
+  const { socket } = req
+  const onSocketClose = () => {
+    clientDisconnected = true
+    logger.debug(
+      `🔌 [Queue] Socket closed during queue wait for API key ${apiKeyId}, requestId: ${requestId}`
+    )
+  }
+
+  if (socket) {
+    socket.once('close', onSocketClose)
+  }
+
+  // 检查 socket 是否在监听器注册前已被销毁（边界情况）
+  if (socket?.destroyed) {
+    clientDisconnected = true
+  }
+
+  const startTime = Date.now()
+  let pollInterval = pollIntervalMs
+  let redisFailCount = 0
+  // 优先使用配置中的值，否则使用默认值
+  const maxRedisFailCount = configMaxRedisFailCount || QUEUE_POLLING_CONFIG.maxRedisFailCount
+
+  try {
+    while (Date.now() - startTime < timeoutMs) {
+      // 检测客户端是否断开（双重检查：事件标记 + socket 状态）
+      // socket.destroyed 是同步检查，确保即使事件处理有延迟也能及时检测
+      if (clientDisconnected || socket?.destroyed) {
+        redis
+          .incrConcurrencyQueueStats(apiKeyId, 'cancelled')
+          .catch((e) => logger.warn('Failed to record cancelled stat:', e))
+        return {
+          acquired: false,
+          reason: 'client_disconnected',
+          waitTimeMs: Date.now() - startTime
+        }
+      }
+
+      // 尝试获取槽位（先占后检查）
+      try {
+        const count = await redis.incrConcurrency(apiKeyId, requestId, leaseSeconds)
+        redisFailCount = 0 // 重置失败计数
+
+        if (count <= concurrencyLimit) {
+          // 成功获取槽位！
+          const waitTimeMs = Date.now() - startTime
+
+          // 槽位所有权转移说明：
+          // 1. 此时槽位已通过 incrConcurrency 获取
+          // 2. 先标记 internalSlotAcquired = true，确保异常时 finally 块能清理
+          // 3. 统计操作完成后，清除标记并返回，所有权转移给调用方
+          // 4. 调用方（authenticateApiKey）负责在请求结束时释放槽位
+
+          // 标记槽位已获取（用于异常时 finally 块清理）
+          internalSlotAcquired = true
+
+          // 记录统计（非阻塞，fire-and-forget 模式）
+          // ⚠️ 设计说明：
+          // - 故意不 await 这些 Promise，因为统计记录不应阻塞请求处理
+          // - 每个 Promise 都有独立的 .catch()，确保单个失败不影响其他
+          // - 外层 .catch() 是防御性措施，处理 Promise.all 本身的异常
+          // - 即使统计记录在函数返回后才完成/失败，也是安全的（仅日志记录）
+          // - 统计数据丢失可接受，不影响核心业务逻辑
+          Promise.all([
+            redis
+              .recordQueueWaitTime(apiKeyId, waitTimeMs)
+              .catch((e) => logger.warn('Failed to record queue wait time:', e)),
+            redis
+              .recordGlobalQueueWaitTime(waitTimeMs)
+              .catch((e) => logger.warn('Failed to record global wait time:', e)),
+            redis
+              .incrConcurrencyQueueStats(apiKeyId, 'success')
+              .catch((e) => logger.warn('Failed to increment success stats:', e))
+          ]).catch((e) => logger.warn('Failed to record queue stats batch:', e))
+
+          // 成功返回前清除标记（所有权转移给调用方，由其负责释放）
+          internalSlotAcquired = false
+          return { acquired: true, waitTimeMs }
+        }
+
+        // 超限，释放槽位继续等待
+        try {
+          await redis.decrConcurrency(apiKeyId, requestId)
+        } catch (decrError) {
+          // 释放失败时记录警告但继续轮询
+          // 下次 incrConcurrency 会自然覆盖同一 requestId 的条目
+          logger.warn(
+            `Failed to release slot during polling for ${apiKeyId}, will retry:`,
+            decrError
+          )
+        }
+      } catch (redisError) {
+        redisFailCount++
+        logger.error(
+          `Redis error in queue polling (${redisFailCount}/${maxRedisFailCount}):`,
+          redisError
+        )
+
+        if (redisFailCount >= maxRedisFailCount) {
+          // 连续 Redis 失败，放弃排队
+          return {
+            acquired: false,
+            reason: 'redis_error',
+            waitTimeMs: Date.now() - startTime
+          }
+        }
+      }
+
+      // 指数退避等待
+      await sleep(pollInterval)
+
+      // 计算下一次轮询间隔（指数退避 + 抖动）
+      // 1. 先应用指数退避
+      let nextInterval = pollInterval * backoffFactor
+      // 2. 添加抖动防止惊群效应（±jitterRatio 范围内的随机偏移）
+      //    抖动范围：[-jitterRatio, +jitterRatio]，例如 jitterRatio=0.2 时为 ±20%
+      //    这是预期行为：负抖动可使间隔略微缩短，正抖动可使间隔略微延长
+      //    目的是分散多个等待者的轮询时间点，避免同时请求 Redis
+      const jitter = nextInterval * jitterRatio * (Math.random() * 2 - 1)
+      nextInterval = nextInterval + jitter
+      // 3. 确保在合理范围内：最小 1ms，最大 maxPollIntervalMs
+      //    Math.max(1, ...) 保证即使负抖动也不会产生 ≤0 的间隔
+      pollInterval = Math.max(1, Math.min(nextInterval, maxPollIntervalMs))
+    }
+
+    // 超时
+    redis
+      .incrConcurrencyQueueStats(apiKeyId, 'timeout')
+      .catch((e) => logger.warn('Failed to record timeout stat:', e))
+    return { acquired: false, reason: 'timeout', waitTimeMs: Date.now() - startTime }
+  } finally {
+    // 确保清理：
+    // 1. 减少排队计数（排队计数在调用方已增加，这里负责减少）
+    try {
+      await redis.decrConcurrencyQueue(apiKeyId)
+    } catch (cleanupError) {
+      // 清理失败记录错误（可能导致计数泄漏，但有 TTL 保护）
+      logger.error(
+        `Failed to decrement queue count in finally block for ${apiKeyId}:`,
+        cleanupError
+      )
+    }
+
+    // 2. 如果内部获取了槽位但未正常返回（异常路径），释放槽位
+    if (internalSlotAcquired) {
+      try {
+        await redis.decrConcurrency(apiKeyId, requestId)
+        logger.warn(
+          `⚠️ Released orphaned concurrency slot in finally block for ${apiKeyId}, requestId: ${requestId}`
+        )
+      } catch (slotCleanupError) {
+        logger.error(
+          `Failed to release orphaned concurrency slot for ${apiKeyId}:`,
+          slotCleanupError
+        )
+      }
+    }
+
+    // 清理 socket 事件监听器
+    if (socket) {
+      socket.removeListener('close', onSocketClose)
+    }
+  }
+}
+
 // 🔑 API Key验证中间件（优化版）
 const authenticateApiKey = async (req, res, next) => {
   const startTime = Date.now()
+  let authErrored = false
+  let concurrencyCleanup = null
+  let hasConcurrencySlot = false
 
   try {
     // 安全提取API Key，支持多种格式（包括Gemini CLI支持）
@@ -201,6 +513,53 @@ const authenticateApiKey = async (req, res, next) => {
       )
     }
 
+    // 🔒 检查全局 Claude Code 限制（与 API Key 级别是 OR 逻辑）
+    // 仅对 Claude 服务端点生效 (/api/v1/messages 和 /claude/v1/messages)
+    if (!skipKeyRestrictions) {
+      const normalizedPath = (req.originalUrl || req.path || '').toLowerCase()
+      const isClaudeMessagesEndpoint =
+        normalizedPath.includes('/v1/messages') &&
+        (normalizedPath.startsWith('/api') || normalizedPath.startsWith('/claude'))
+
+      if (isClaudeMessagesEndpoint) {
+        try {
+          const globalClaudeCodeOnly = await claudeRelayConfigService.isClaudeCodeOnlyEnabled()
+
+          // API Key 级别的 Claude Code 限制
+          const keyClaudeCodeOnly =
+            validation.keyData.enableClientRestriction &&
+            Array.isArray(validation.keyData.allowedClients) &&
+            validation.keyData.allowedClients.length === 1 &&
+            validation.keyData.allowedClients.includes('claude_code')
+
+          // OR 逻辑：全局开启 或 API Key 级别限制为仅 claude_code
+          if (globalClaudeCodeOnly || keyClaudeCodeOnly) {
+            const isClaudeCode = ClaudeCodeValidator.validate(req)
+
+            if (!isClaudeCode) {
+              const clientIP = req.ip || req.connection?.remoteAddress || 'unknown'
+              logger.api(
+                `❌ Claude Code client validation failed (global: ${globalClaudeCodeOnly}, key: ${keyClaudeCodeOnly}) from ${clientIP}`
+              )
+              return res.status(403).json({
+                error: {
+                  type: 'client_validation_error',
+                  message: 'This endpoint only accepts requests from Claude Code CLI'
+                }
+              })
+            }
+
+            logger.api(
+              `✅ Claude Code client validated (global: ${globalClaudeCodeOnly}, key: ${keyClaudeCodeOnly})`
+            )
+          }
+        } catch (error) {
+          logger.error('❌ Error checking Claude Code restriction:', error)
+          // 配置服务出错时不阻断请求
+        }
+      }
+    }
+
     // 检查并发限制
     const concurrencyLimit = validation.keyData.concurrencyLimit || 0
     if (!skipKeyRestrictions && concurrencyLimit > 0) {
@@ -216,29 +575,346 @@ const authenticateApiKey = async (req, res, next) => {
       }
       const requestId = uuidv4()
 
+      // ⚠️ 优化后的 Connection: close 设置策略
+      // 问题背景：HTTP Keep-Alive 使多个请求共用同一个 TCP 连接
+      // 当第一个请求正在处理，第二个请求进入排队时，它们共用同一个 socket
+      // 如果客户端超时关闭连接，两个请求都会受影响
+      // 优化方案：只有在请求实际进入排队时才设置 Connection: close
+      // 未排队的请求保持 Keep-Alive，避免不必要的 TCP 握手开销
+      // 详见 design.md Decision 2: Connection: close 设置时机
+      // 注意：Connection: close 将在下方代码实际进入排队时设置（第 637 行左右）
+
+      // ============================================================
+      // 🔒 并发槽位状态管理说明
+      // ============================================================
+      // 此函数中有两个关键状态变量：
+      // - hasConcurrencySlot: 当前是否持有并发槽位
+      // - concurrencyCleanup: 错误时调用的清理函数
+      //
+      // 状态转换流程：
+      // 1. incrConcurrency 成功 → hasConcurrencySlot=true, 设置临时清理函数
+      // 2. 若超限 → 释放槽位，hasConcurrencySlot=false, concurrencyCleanup=null
+      // 3. 若排队成功 → hasConcurrencySlot=true, 升级为完整清理函数（含 interval 清理）
+      // 4. 请求结束（res.close/req.close）→ 调用 decrementConcurrency 释放
+      // 5. 认证错误 → finally 块调用 concurrencyCleanup 释放
+      //
+      // 为什么需要两种清理函数？
+      // - 临时清理：在排队/认证过程中出错时使用，只释放槽位
+      // - 完整清理：请求正常开始后使用，还需清理 leaseRenewInterval
+      // ============================================================
+      const setTemporaryConcurrencyCleanup = () => {
+        concurrencyCleanup = async () => {
+          if (!hasConcurrencySlot) {
+            return
+          }
+          hasConcurrencySlot = false
+          try {
+            await redis.decrConcurrency(validation.keyData.id, requestId)
+          } catch (cleanupError) {
+            logger.error(
+              `Failed to decrement concurrency after auth error for key ${validation.keyData.id}:`,
+              cleanupError
+            )
+          }
+        }
+      }
+
       const currentConcurrency = await redis.incrConcurrency(
         validation.keyData.id,
         requestId,
         leaseSeconds
       )
+      hasConcurrencySlot = true
+      setTemporaryConcurrencyCleanup()
       logger.api(
         `📈 Incremented concurrency for key: ${validation.keyData.id} (${validation.keyData.name}), current: ${currentConcurrency}, limit: ${concurrencyLimit}`
       )
 
       if (currentConcurrency > concurrencyLimit) {
-        // 如果超过限制，立即减少计数
-        await redis.decrConcurrency(validation.keyData.id, requestId)
-        logger.security(
-          `🚦 Concurrency limit exceeded for key: ${validation.keyData.id} (${
-            validation.keyData.name
-          }), current: ${currentConcurrency - 1}, limit: ${concurrencyLimit}`
+        // 1. 先释放刚占用的槽位
+        try {
+          await redis.decrConcurrency(validation.keyData.id, requestId)
+        } catch (error) {
+          logger.error(
+            `Failed to decrement concurrency after limit exceeded for key ${validation.keyData.id}:`,
+            error
+          )
+        }
+        hasConcurrencySlot = false
+        concurrencyCleanup = null
+
+        // 2. 获取排队配置
+        const queueConfig = await claudeRelayConfigService.getConfig()
+
+        // 3. 排队功能未启用，直接返回 429（保持现有行为）
+        if (!queueConfig.concurrentRequestQueueEnabled) {
+          logger.security(
+            `🚦 Concurrency limit exceeded for key: ${validation.keyData.id} (${
+              validation.keyData.name
+            }), current: ${currentConcurrency - 1}, limit: ${concurrencyLimit}`
+          )
+          // 建议客户端在短暂延迟后重试（并发场景下通常很快会有槽位释放）
+          res.set('Retry-After', '1')
+          return res.status(429).json({
+            error: 'Concurrency limit exceeded',
+            message: `Too many concurrent requests. Limit: ${concurrencyLimit} concurrent requests`,
+            currentConcurrency: currentConcurrency - 1,
+            concurrencyLimit
+          })
+        }
+
+        // 4. 计算最大排队数
+        const maxQueueSize = Math.max(
+          concurrencyLimit * queueConfig.concurrentRequestQueueMaxSizeMultiplier,
+          queueConfig.concurrentRequestQueueMaxSize
         )
-        return res.status(429).json({
-          error: 'Concurrency limit exceeded',
-          message: `Too many concurrent requests. Limit: ${concurrencyLimit} concurrent requests`,
-          currentConcurrency: currentConcurrency - 1,
-          concurrencyLimit
-        })
+
+        // 4.5 排队健康检查：过载时快速失败
+        // 详见 design.md Decision 7: 排队健康检查与快速失败
+        const overloadCheck = await shouldRejectDueToOverload(
+          validation.keyData.id,
+          queueConfig.concurrentRequestQueueTimeoutMs,
+          queueConfig,
+          maxQueueSize
+        )
+        if (overloadCheck.reject) {
+          // 使用健康检查返回的当前排队数，避免重复调用 Redis
+          const currentQueueCount = overloadCheck.currentQueueCount || 0
+          logger.api(
+            `🚨 Queue overloaded for key: ${validation.keyData.id} (${validation.keyData.name}), ` +
+              `P90=${overloadCheck.estimatedWaitMs}ms, timeout=${overloadCheck.timeoutMs}ms, ` +
+              `threshold=${overloadCheck.threshold}, samples=${overloadCheck.sampleCount}, ` +
+              `concurrency=${concurrencyLimit}, queue=${currentQueueCount}/${maxQueueSize}`
+          )
+          // 记录被拒绝的过载统计
+          redis
+            .incrConcurrencyQueueStats(validation.keyData.id, 'rejected_overload')
+            .catch((e) => logger.warn('Failed to record rejected_overload stat:', e))
+          // 返回 429 + Retry-After，让客户端稍后重试
+          const retryAfterSeconds = 30
+          res.set('Retry-After', String(retryAfterSeconds))
+          return res.status(429).json({
+            error: 'Queue overloaded',
+            message: `Queue is overloaded. Estimated wait time (${overloadCheck.estimatedWaitMs}ms) exceeds threshold. Limit: ${concurrencyLimit} concurrent requests, queue: ${currentQueueCount}/${maxQueueSize}. Please retry later.`,
+            currentConcurrency: concurrencyLimit,
+            concurrencyLimit,
+            queueCount: currentQueueCount,
+            maxQueueSize,
+            estimatedWaitMs: overloadCheck.estimatedWaitMs,
+            timeoutMs: overloadCheck.timeoutMs,
+            queueTimeoutMs: queueConfig.concurrentRequestQueueTimeoutMs,
+            retryAfterSeconds
+          })
+        }
+
+        // 5. 尝试进入排队（原子操作：先增加再检查，避免竞态条件）
+        let queueIncremented = false
+        try {
+          const newQueueCount = await redis.incrConcurrencyQueue(
+            validation.keyData.id,
+            queueConfig.concurrentRequestQueueTimeoutMs
+          )
+          queueIncremented = true
+
+          if (newQueueCount > maxQueueSize) {
+            // 超过最大排队数，立即释放并返回 429
+            await redis.decrConcurrencyQueue(validation.keyData.id)
+            queueIncremented = false
+            logger.api(
+              `🚦 Concurrency queue full for key: ${validation.keyData.id} (${validation.keyData.name}), ` +
+                `queue: ${newQueueCount - 1}, maxQueue: ${maxQueueSize}`
+            )
+            // 队列已满，建议客户端在排队超时时间后重试
+            const retryAfterSeconds = Math.ceil(queueConfig.concurrentRequestQueueTimeoutMs / 1000)
+            res.set('Retry-After', String(retryAfterSeconds))
+            return res.status(429).json({
+              error: 'Concurrency queue full',
+              message: `Too many requests waiting in queue. Limit: ${concurrencyLimit} concurrent requests, queue: ${newQueueCount - 1}/${maxQueueSize}, timeout: ${retryAfterSeconds}s`,
+              currentConcurrency: concurrencyLimit,
+              concurrencyLimit,
+              queueCount: newQueueCount - 1,
+              maxQueueSize,
+              queueTimeoutMs: queueConfig.concurrentRequestQueueTimeoutMs,
+              retryAfterSeconds
+            })
+          }
+
+          // 6. 已成功进入排队，记录统计并开始等待槽位
+          logger.api(
+            `⏳ Request entering queue for key: ${validation.keyData.id} (${validation.keyData.name}), ` +
+              `queue position: ${newQueueCount}`
+          )
+          redis
+            .incrConcurrencyQueueStats(validation.keyData.id, 'entered')
+            .catch((e) => logger.warn('Failed to record entered stat:', e))
+
+          // ⚠️ 仅在请求实际进入排队时设置 Connection: close
+          // 详见 design.md Decision 2: Connection: close 设置时机
+          // 未排队的请求保持 Keep-Alive，避免不必要的 TCP 握手开销
+          if (!res.headersSent) {
+            res.setHeader('Connection', 'close')
+            logger.api(
+              `🔌 [Queue] Set Connection: close for queued request, key: ${validation.keyData.id}`
+            )
+          }
+
+          // ⚠️ 记录排队开始时的 socket 标识，用于排队完成后验证
+          // 问题背景：HTTP Keep-Alive 连接复用时，长时间排队可能导致 socket 被其他请求使用
+          // 验证方法：使用 UUID token + socket 对象引用双重验证
+          // 详见 design.md Decision 1: Socket 身份验证机制
+          req._crService = req._crService || {}
+          req._crService.queueToken = uuidv4()
+          req._crService.originalSocket = req.socket
+          req._crService.startTime = Date.now()
+          const savedToken = req._crService.queueToken
+          const savedSocket = req._crService.originalSocket
+
+          // ⚠️ 重要：在调用前将 queueIncremented 设为 false
+          // 因为 waitForConcurrencySlot 的 finally 块会负责清理排队计数
+          // 如果在调用后设置，当 waitForConcurrencySlot 抛出异常时
+          // 外层 catch 块会重复减少计数（finally 已经减过一次）
+          queueIncremented = false
+
+          const slot = await waitForConcurrencySlot(req, res, validation.keyData.id, {
+            concurrencyLimit,
+            requestId,
+            leaseSeconds,
+            timeoutMs: queueConfig.concurrentRequestQueueTimeoutMs,
+            pollIntervalMs: QUEUE_POLLING_CONFIG.pollIntervalMs,
+            maxPollIntervalMs: QUEUE_POLLING_CONFIG.maxPollIntervalMs,
+            backoffFactor: QUEUE_POLLING_CONFIG.backoffFactor,
+            jitterRatio: QUEUE_POLLING_CONFIG.jitterRatio,
+            maxRedisFailCount: queueConfig.concurrentRequestQueueMaxRedisFailCount
+          })
+
+          // 7. 处理排队结果
+          if (!slot.acquired) {
+            if (slot.reason === 'client_disconnected') {
+              // 客户端已断开，不返回响应（连接已关闭）
+              logger.api(
+                `🔌 Client disconnected while queuing for key: ${validation.keyData.id} (${validation.keyData.name})`
+              )
+              return
+            }
+
+            if (slot.reason === 'redis_error') {
+              // Redis 连续失败，返回 503
+              logger.error(
+                `❌ Redis error during queue wait for key: ${validation.keyData.id} (${validation.keyData.name})`
+              )
+              return res.status(503).json({
+                error: 'Service temporarily unavailable',
+                message: 'Failed to acquire concurrency slot due to internal error'
+              })
+            }
+            // 排队超时（使用 api 级别，与其他排队日志保持一致）
+            logger.api(
+              `⏰ Queue timeout for key: ${validation.keyData.id} (${validation.keyData.name}), waited: ${slot.waitTimeMs}ms`
+            )
+            // 已等待超时，建议客户端稍后重试
+            // ⚠️ Retry-After 策略优化：
+            // - 请求已经等了完整的 timeout 时间，说明系统负载较高
+            // - 过早重试（如固定 5 秒）会加剧拥塞，导致更多超时
+            // - 合理策略：使用 timeout 时间的一半作为重试间隔
+            // - 最小值 5 秒，最大值 30 秒，避免极端情况
+            const timeoutSeconds = Math.ceil(queueConfig.concurrentRequestQueueTimeoutMs / 1000)
+            const retryAfterSeconds = Math.max(5, Math.min(30, Math.ceil(timeoutSeconds / 2)))
+            res.set('Retry-After', String(retryAfterSeconds))
+            return res.status(429).json({
+              error: 'Queue timeout',
+              message: `Request timed out waiting for concurrency slot. Limit: ${concurrencyLimit} concurrent requests, maxQueue: ${maxQueueSize}, Queue timeout: ${timeoutSeconds}s, waited: ${slot.waitTimeMs}ms`,
+              currentConcurrency: concurrencyLimit,
+              concurrencyLimit,
+              maxQueueSize,
+              queueTimeoutMs: queueConfig.concurrentRequestQueueTimeoutMs,
+              waitTimeMs: slot.waitTimeMs,
+              retryAfterSeconds
+            })
+          }
+
+          // 8. 排队成功，slot.acquired 表示已在 waitForConcurrencySlot 中获取到槽位
+          logger.api(
+            `✅ Queue wait completed for key: ${validation.keyData.id} (${validation.keyData.name}), ` +
+              `waited: ${slot.waitTimeMs}ms`
+          )
+          hasConcurrencySlot = true
+          setTemporaryConcurrencyCleanup()
+
+          // 9. ⚠️ 关键检查：排队等待结束后，验证客户端是否还在等待响应
+          // 长时间排队后，客户端可能在应用层已放弃（如 Claude Code 的超时机制），
+          // 但 TCP 连接仍然存活。此时继续处理请求是浪费资源。
+          // 注意：如果发送了心跳，headersSent 会是 true，但这是正常的
+          const postQueueSocket = req.socket
+          // 只检查连接是否真正断开（destroyed/writableEnded/socketDestroyed）
+          // headersSent 在心跳场景下是正常的，不应该作为放弃的依据
+          if (res.destroyed || res.writableEnded || postQueueSocket?.destroyed) {
+            logger.warn(
+              `⚠️ Client no longer waiting after queue for key: ${validation.keyData.id} (${validation.keyData.name}), ` +
+                `waited: ${slot.waitTimeMs}ms | destroyed: ${res.destroyed}, ` +
+                `writableEnded: ${res.writableEnded}, socketDestroyed: ${postQueueSocket?.destroyed}`
+            )
+            // 释放刚获取的槽位
+            hasConcurrencySlot = false
+            await redis
+              .decrConcurrency(validation.keyData.id, requestId)
+              .catch((e) => logger.error('Failed to release slot after client abandoned:', e))
+            // 不返回响应（客户端已不在等待）
+            return
+          }
+
+          // 10. ⚠️ 关键检查：验证 socket 身份是否改变
+          // HTTP Keep-Alive 连接复用可能导致排队期间 socket 被其他请求使用
+          // 验证方法：UUID token + socket 对象引用双重验证
+          // 详见 design.md Decision 1: Socket 身份验证机制
+          const queueData = req._crService
+          const socketIdentityChanged =
+            !queueData ||
+            queueData.queueToken !== savedToken ||
+            queueData.originalSocket !== savedSocket
+
+          if (socketIdentityChanged) {
+            logger.error(
+              `❌ [Queue] Socket identity changed during queue wait! ` +
+                `key: ${validation.keyData.id} (${validation.keyData.name}), ` +
+                `waited: ${slot.waitTimeMs}ms | ` +
+                `tokenMatch: ${queueData?.queueToken === savedToken}, ` +
+                `socketMatch: ${queueData?.originalSocket === savedSocket}`
+            )
+            // 释放刚获取的槽位
+            hasConcurrencySlot = false
+            await redis
+              .decrConcurrency(validation.keyData.id, requestId)
+              .catch((e) => logger.error('Failed to release slot after socket identity change:', e))
+            // 记录 socket_changed 统计
+            redis
+              .incrConcurrencyQueueStats(validation.keyData.id, 'socket_changed')
+              .catch((e) => logger.warn('Failed to record socket_changed stat:', e))
+            // 不返回响应（socket 已被其他请求使用）
+            return
+          }
+        } catch (queueError) {
+          // 异常时清理资源，防止泄漏
+          // 1. 清理排队计数（如果还没被 waitForConcurrencySlot 的 finally 清理）
+          if (queueIncremented) {
+            await redis
+              .decrConcurrencyQueue(validation.keyData.id)
+              .catch((e) => logger.error('Failed to cleanup queue count after error:', e))
+          }
+
+          // 2. 防御性清理：如果 waitForConcurrencySlot 内部获取了槽位但在返回前异常
+          //    虽然这种情况极少发生（统计记录的异常会被内部捕获），但为了安全起见
+          //    尝试释放可能已获取的槽位。decrConcurrency 使用 ZREM，即使成员不存在也安全
+          if (hasConcurrencySlot) {
+            hasConcurrencySlot = false
+            await redis
+              .decrConcurrency(validation.keyData.id, requestId)
+              .catch((e) =>
+                logger.error('Failed to cleanup concurrency slot after queue error:', e)
+              )
+          }
+
+          throw queueError
+        }
       }
 
       const renewIntervalMs =
@@ -249,7 +925,38 @@ const authenticateApiKey = async (req, res, next) => {
       let leaseRenewInterval = null
 
       if (renewIntervalMs > 0) {
+        // 🔴 关键修复：添加最大刷新次数限制，防止租约永不过期
+        // 默认最大生存时间为 10 分钟，可通过环境变量配置
+        const maxLifetimeMinutes = parseInt(process.env.CONCURRENCY_MAX_LIFETIME_MINUTES) || 10
+        const maxRefreshCount = Math.ceil((maxLifetimeMinutes * 60 * 1000) / renewIntervalMs)
+        let refreshCount = 0
+
         leaseRenewInterval = setInterval(() => {
+          refreshCount++
+
+          // 超过最大刷新次数，强制停止并清理
+          if (refreshCount > maxRefreshCount) {
+            logger.warn(
+              `⚠️ Lease refresh exceeded max count (${maxRefreshCount}) for key ${validation.keyData.id} (${validation.keyData.name}), forcing cleanup after ${maxLifetimeMinutes} minutes`
+            )
+            // 清理定时器
+            if (leaseRenewInterval) {
+              clearInterval(leaseRenewInterval)
+              leaseRenewInterval = null
+            }
+            // 强制减少并发计数（如果还没减少）
+            if (!concurrencyDecremented) {
+              concurrencyDecremented = true
+              redis.decrConcurrency(validation.keyData.id, requestId).catch((error) => {
+                logger.error(
+                  `Failed to decrement concurrency after max refresh for key ${validation.keyData.id}:`,
+                  error
+                )
+              })
+            }
+            return
+          }
+
           redis
             .refreshConcurrencyLease(validation.keyData.id, requestId, leaseSeconds)
             .catch((error) => {
@@ -268,6 +975,7 @@ const authenticateApiKey = async (req, res, next) => {
       const decrementConcurrency = async () => {
         if (!concurrencyDecremented) {
           concurrencyDecremented = true
+          hasConcurrencySlot = false
           if (leaseRenewInterval) {
             clearInterval(leaseRenewInterval)
             leaseRenewInterval = null
@@ -282,6 +990,11 @@ const authenticateApiKey = async (req, res, next) => {
           }
         }
       }
+      // 升级为完整清理函数（包含 leaseRenewInterval 清理逻辑）
+      // 此时请求已通过认证，后续由 res.close/req.close 事件触发清理
+      if (hasConcurrencySlot) {
+        concurrencyCleanup = decrementConcurrency
+      }
 
       // 监听最可靠的事件（避免重复监听）
       // res.on('close') 是最可靠的，会在连接关闭时触发
@@ -607,6 +1320,7 @@ const authenticateApiKey = async (req, res, next) => {
 
     return next()
   } catch (error) {
+    authErrored = true
     const authDuration = Date.now() - startTime
     logger.error(`❌ Authentication middleware error (${authDuration}ms):`, {
       error: error.message,
@@ -620,6 +1334,14 @@ const authenticateApiKey = async (req, res, next) => {
       error: 'Authentication error',
       message: 'Internal server error during authentication'
     })
+  } finally {
+    if (authErrored && typeof concurrencyCleanup === 'function') {
+      try {
+        await concurrencyCleanup()
+      } catch (cleanupError) {
+        logger.error('Failed to cleanup concurrency after auth error:', cleanupError)
+      }
+    }
   }
 }
 
@@ -667,6 +1389,18 @@ const authenticateAdmin = async (req, res, next) => {
       })
     }
 
+    // 🔒 安全修复：验证会话必须字段（防止伪造会话绕过认证）
+    if (!adminSession.username || !adminSession.loginTime) {
+      logger.security(
+        `🔒 Corrupted admin session from ${req.ip || 'unknown'} - missing required fields (username: ${!!adminSession.username}, loginTime: ${!!adminSession.loginTime})`
+      )
+      await redis.deleteSession(token) // 清理无效/伪造的会话
+      return res.status(401).json({
+        error: 'Invalid session',
+        message: 'Session data corrupted or incomplete'
+      })
+    }
+
     // 检查会话活跃性（可选：检查最后活动时间）
     const now = new Date()
     const lastActivity = new Date(adminSession.lastActivity || adminSession.loginTime)
@@ -700,7 +1434,6 @@ const authenticateAdmin = async (req, res, next) => {
 
     // 设置管理员信息（只包含必要信息）
     req.admin = {
-      id: adminSession.adminId || 'admin',
       username: adminSession.username,
       sessionId: token,
       loginTime: adminSession.loginTime
@@ -833,17 +1566,25 @@ const authenticateUserOrAdmin = async (req, res, next) => {
       try {
         const adminSession = await redis.getSession(adminToken)
         if (adminSession && Object.keys(adminSession).length > 0) {
-          req.admin = {
-            id: adminSession.adminId || 'admin',
-            username: adminSession.username,
-            sessionId: adminToken,
-            loginTime: adminSession.loginTime
-          }
-          req.userType = 'admin'
+          // 🔒 安全修复：验证会话必须字段（与 authenticateAdmin 保持一致）
+          if (!adminSession.username || !adminSession.loginTime) {
+            logger.security(
+              `🔒 Corrupted admin session in authenticateUserOrAdmin from ${req.ip || 'unknown'} - missing required fields (username: ${!!adminSession.username}, loginTime: ${!!adminSession.loginTime})`
+            )
+            await redis.deleteSession(adminToken) // 清理无效/伪造的会话
+            // 不返回 401，继续尝试用户认证
+          } else {
+            req.admin = {
+              username: adminSession.username,
+              sessionId: adminToken,
+              loginTime: adminSession.loginTime
+            }
+            req.userType = 'admin'
 
-          const authDuration = Date.now() - startTime
-          logger.security(`🔐 Admin authenticated: ${adminSession.username} in ${authDuration}ms`)
-          return next()
+            const authDuration = Date.now() - startTime
+            logger.security(`🔐 Admin authenticated: ${adminSession.username} in ${authDuration}ms`)
+            return next()
+          }
         }
       } catch (error) {
         logger.debug('Admin authentication failed, trying user authentication:', error.message)
@@ -1022,9 +1763,13 @@ const requestLogger = (req, res, next) => {
   const referer = req.get('Referer') || 'none'
 
   // 记录请求开始
+  const isDebugRoute = req.originalUrl.includes('event_logging')
   if (req.originalUrl !== '/health') {
-    // 避免健康检查日志过多
-    logger.info(`▶️ [${requestId}] ${req.method} ${req.originalUrl} | IP: ${clientIP}`)
+    if (isDebugRoute) {
+      logger.debug(`▶️ [${requestId}] ${req.method} ${req.originalUrl} | IP: ${clientIP}`)
+    } else {
+      logger.info(`▶️ [${requestId}] ${req.method} ${req.originalUrl} | IP: ${clientIP}`)
+    }
   }
 
   res.on('finish', () => {
@@ -1056,7 +1801,14 @@ const requestLogger = (req, res, next) => {
         logMetadata
       )
     } else if (req.originalUrl !== '/health') {
-      logger.request(req.method, req.originalUrl, res.statusCode, duration, logMetadata)
+      if (isDebugRoute) {
+        logger.debug(
+          `🟢 ${req.method} ${req.originalUrl} - ${res.statusCode} (${duration}ms)`,
+          logMetadata
+        )
+      } else {
+        logger.request(req.method, req.originalUrl, res.statusCode, duration, logMetadata)
+      }
     }
 
     // API Key相关日志
@@ -1298,7 +2050,8 @@ const globalRateLimit = async (req, res, next) =>
 
 // 📊 请求大小限制中间件
 const requestSizeLimit = (req, res, next) => {
-  const maxSize = 60 * 1024 * 1024 // 60MB
+  const MAX_SIZE_MB = parseInt(process.env.REQUEST_MAX_SIZE_MB || '100', 10)
+  const maxSize = MAX_SIZE_MB * 1024 * 1024
   const contentLength = parseInt(req.headers['content-length'] || '0')
 
   if (contentLength > maxSize) {
@@ -1306,7 +2059,7 @@ const requestSizeLimit = (req, res, next) => {
     return res.status(413).json({
       error: 'Payload Too Large',
       message: 'Request body size exceeds limit',
-      limit: '10MB'
+      limit: `${MAX_SIZE_MB}MB`
     })
   }
 

src/routes/admin/accountBalance.js

@@ -0,0 +1,214 @@
+const express = require('express')
+const { authenticateAdmin } = require('../../middleware/auth')
+const logger = require('../../utils/logger')
+const accountBalanceService = require('../../services/accountBalanceService')
+const balanceScriptService = require('../../services/balanceScriptService')
+const { isBalanceScriptEnabled } = require('../../utils/featureFlags')
+
+const router = express.Router()
+
+const ensureValidPlatform = (rawPlatform) => {
+  const normalized = accountBalanceService.normalizePlatform(rawPlatform)
+  if (!normalized) {
+    return { ok: false, status: 400, error: '缺少 platform 参数' }
+  }
+
+  const supported = accountBalanceService.getSupportedPlatforms()
+  if (!supported.includes(normalized)) {
+    return { ok: false, status: 400, error: `不支持的平台: ${normalized}` }
+  }
+
+  return { ok: true, platform: normalized }
+}
+
+// 1) 获取账户余额（默认本地统计优先，可选触发 Provider）
+// GET /admin/accounts/:accountId/balance?platform=xxx&queryApi=false
+router.get('/accounts/:accountId/balance', authenticateAdmin, async (req, res) => {
+  try {
+    const { accountId } = req.params
+    const { platform, queryApi } = req.query
+
+    const valid = ensureValidPlatform(platform)
+    if (!valid.ok) {
+      return res.status(valid.status).json({ success: false, error: valid.error })
+    }
+
+    const balance = await accountBalanceService.getAccountBalance(accountId, valid.platform, {
+      queryApi
+    })
+
+    if (!balance) {
+      return res.status(404).json({ success: false, error: 'Account not found' })
+    }
+
+    return res.json(balance)
+  } catch (error) {
+    logger.error('获取账户余额失败', error)
+    return res.status(500).json({ success: false, error: error.message })
+  }
+})
+
+// 2) 强制刷新账户余额（强制触发查询：优先脚本；Provider 仅为降级）
+// POST /admin/accounts/:accountId/balance/refresh
+// Body: { platform: 'xxx' }
+router.post('/accounts/:accountId/balance/refresh', authenticateAdmin, async (req, res) => {
+  try {
+    const { accountId } = req.params
+    const { platform } = req.body || {}
+
+    const valid = ensureValidPlatform(platform)
+    if (!valid.ok) {
+      return res.status(valid.status).json({ success: false, error: valid.error })
+    }
+
+    logger.info(`手动刷新余额: ${valid.platform}:${accountId}`)
+
+    const balance = await accountBalanceService.refreshAccountBalance(accountId, valid.platform)
+    if (!balance) {
+      return res.status(404).json({ success: false, error: 'Account not found' })
+    }
+
+    return res.json(balance)
+  } catch (error) {
+    logger.error('刷新账户余额失败', error)
+    return res.status(500).json({ success: false, error: error.message })
+  }
+})
+
+// 3) 批量获取平台所有账户余额
+// GET /admin/accounts/balance/platform/:platform?queryApi=false
+router.get('/accounts/balance/platform/:platform', authenticateAdmin, async (req, res) => {
+  try {
+    const { platform } = req.params
+    const { queryApi } = req.query
+
+    const valid = ensureValidPlatform(platform)
+    if (!valid.ok) {
+      return res.status(valid.status).json({ success: false, error: valid.error })
+    }
+
+    const balances = await accountBalanceService.getAllAccountsBalance(valid.platform, { queryApi })
+
+    return res.json({ success: true, data: balances })
+  } catch (error) {
+    logger.error('批量获取余额失败', error)
+    return res.status(500).json({ success: false, error: error.message })
+  }
+})
+
+// 4) 获取余额汇总（Dashboard 用）
+// GET /admin/accounts/balance/summary
+router.get('/accounts/balance/summary', authenticateAdmin, async (req, res) => {
+  try {
+    const summary = await accountBalanceService.getBalanceSummary()
+    return res.json({ success: true, data: summary })
+  } catch (error) {
+    logger.error('获取余额汇总失败', error)
+    return res.status(500).json({ success: false, error: error.message })
+  }
+})
+
+// 5) 清除缓存
+// DELETE /admin/accounts/:accountId/balance/cache?platform=xxx
+router.delete('/accounts/:accountId/balance/cache', authenticateAdmin, async (req, res) => {
+  try {
+    const { accountId } = req.params
+    const { platform } = req.query
+
+    const valid = ensureValidPlatform(platform)
+    if (!valid.ok) {
+      return res.status(valid.status).json({ success: false, error: valid.error })
+    }
+
+    await accountBalanceService.clearCache(accountId, valid.platform)
+
+    return res.json({ success: true, message: '缓存已清除' })
+  } catch (error) {
+    logger.error('清除缓存失败', error)
+    return res.status(500).json({ success: false, error: error.message })
+  }
+})
+
+// 6) 获取/保存/测试余额脚本配置（单账户）
+router.get('/accounts/:accountId/balance/script', authenticateAdmin, async (req, res) => {
+  try {
+    const { accountId } = req.params
+    const { platform } = req.query
+
+    const valid = ensureValidPlatform(platform)
+    if (!valid.ok) {
+      return res.status(valid.status).json({ success: false, error: valid.error })
+    }
+
+    const config = await accountBalanceService.redis.getBalanceScriptConfig(
+      valid.platform,
+      accountId
+    )
+    return res.json({ success: true, data: config || null })
+  } catch (error) {
+    logger.error('获取余额脚本配置失败', error)
+    return res.status(500).json({ success: false, error: error.message })
+  }
+})
+
+router.put('/accounts/:accountId/balance/script', authenticateAdmin, async (req, res) => {
+  try {
+    const { accountId } = req.params
+    const { platform } = req.query
+    const valid = ensureValidPlatform(platform)
+    if (!valid.ok) {
+      return res.status(valid.status).json({ success: false, error: valid.error })
+    }
+
+    const payload = req.body || {}
+    await accountBalanceService.redis.setBalanceScriptConfig(valid.platform, accountId, payload)
+    return res.json({ success: true, data: payload })
+  } catch (error) {
+    logger.error('保存余额脚本配置失败', error)
+    return res.status(500).json({ success: false, error: error.message })
+  }
+})
+
+router.post('/accounts/:accountId/balance/script/test', authenticateAdmin, async (req, res) => {
+  try {
+    const { accountId } = req.params
+    const { platform } = req.query
+    const valid = ensureValidPlatform(platform)
+    if (!valid.ok) {
+      return res.status(valid.status).json({ success: false, error: valid.error })
+    }
+
+    if (!isBalanceScriptEnabled()) {
+      return res.status(403).json({
+        success: false,
+        error: '余额脚本功能已禁用（可通过 BALANCE_SCRIPT_ENABLED=true 启用）'
+      })
+    }
+
+    const payload = req.body || {}
+    const { scriptBody } = payload
+    if (!scriptBody) {
+      return res.status(400).json({ success: false, error: '脚本内容不能为空' })
+    }
+
+    const result = await balanceScriptService.execute({
+      scriptBody,
+      timeoutSeconds: payload.timeoutSeconds || 10,
+      variables: {
+        baseUrl: payload.baseUrl || '',
+        apiKey: payload.apiKey || '',
+        token: payload.token || '',
+        accountId,
+        platform: valid.platform,
+        extra: payload.extra || ''
+      }
+    })
+
+    return res.json({ success: true, data: result })
+  } catch (error) {
+    logger.error('测试余额脚本失败', error)
+    return res.status(400).json({ success: false, error: error.message })
+  }
+})
+
+module.exports = router

src/routes/admin/apiKeys.js

@@ -8,6 +8,43 @@ const config = require('../../../config/config')
 
 const router = express.Router()
 
+// 有效的权限值列表
+const VALID_PERMISSIONS = ['claude', 'gemini', 'openai', 'droid']
+
+/**
+ * 验证权限数组格式
+ * @param {any} permissions - 权限值（可以是数组或其他）
+ * @returns {string|null} - 返回错误消息，null 表示验证通过
+ */
+function validatePermissions(permissions) {
+  // 空值或未定义表示全部服务
+  if (permissions === undefined || permissions === null || permissions === '') {
+    return null
+  }
+  // 兼容旧格式字符串
+  if (typeof permissions === 'string') {
+    if (permissions === 'all' || VALID_PERMISSIONS.includes(permissions)) {
+      return null
+    }
+    return `Invalid permissions value. Must be an array of: ${VALID_PERMISSIONS.join(', ')}`
+  }
+  // 新格式数组
+  if (Array.isArray(permissions)) {
+    // 空数组表示全部服务
+    if (permissions.length === 0) {
+      return null
+    }
+    // 验证数组中的每个值
+    for (const perm of permissions) {
+      if (!VALID_PERMISSIONS.includes(perm)) {
+        return `Invalid permission value "${perm}". Valid values are: ${VALID_PERMISSIONS.join(', ')}`
+      }
+    }
+    return null
+  }
+  return `Permissions must be an array. Valid values are: ${VALID_PERMISSIONS.join(', ')}`
+}
+
 // 👥 用户管理 (用于API Key分配)
 
 // 获取所有用户列表（用于API Key分配）
@@ -103,6 +140,17 @@ router.get('/api-keys/:keyId/cost-debug', authenticateAdmin, async (req, res) =>
   }
 })
 
+// 获取所有被使用过的模型列表
+router.get('/api-keys/used-models', authenticateAdmin, async (req, res) => {
+  try {
+    const models = await redis.getAllUsedModels()
+    return res.json({ success: true, data: models })
+  } catch (error) {
+    logger.error('❌ Failed to get used models:', error)
+    return res.status(500).json({ error: 'Failed to get used models', message: error.message })
+  }
+})
+
 // 获取所有API Keys
 router.get('/api-keys', authenticateAdmin, async (req, res) => {
   try {
@@ -116,6 +164,7 @@ router.get('/api-keys', authenticateAdmin, async (req, res) => {
       // 筛选参数
       tag = '',
       isActive = '',
+      models = '', // 模型筛选（逗号分隔）
       // 排序参数
       sortBy = 'createdAt',
       sortOrder = 'desc',
@@ -127,6 +176,9 @@ router.get('/api-keys', authenticateAdmin, async (req, res) => {
       timeRange = 'all'
     } = req.query
 
+    // 解析模型筛选参数
+    const modelFilter = models ? models.split(',').filter((m) => m.trim()) : []
+
     // 验证分页参数
     const pageNum = Math.max(1, parseInt(page) || 1)
     const pageSizeNum = [10, 20, 50, 100].includes(parseInt(pageSize)) ? parseInt(pageSize) : 20
@@ -217,7 +269,8 @@ router.get('/api-keys', authenticateAdmin, async (req, res) => {
           search,
           searchMode,
           tag,
-          isActive
+          isActive,
+          modelFilter
         })
 
         costSortStatus = {
@@ -250,7 +303,8 @@ router.get('/api-keys', authenticateAdmin, async (req, res) => {
           search,
           searchMode,
           tag,
-          isActive
+          isActive,
+          modelFilter
         })
 
         costSortStatus.isRealTimeCalculation = false
@@ -265,7 +319,8 @@ router.get('/api-keys', authenticateAdmin, async (req, res) => {
         tag,
         isActive,
         sortBy: validSortBy,
-        sortOrder: validSortOrder
+        sortOrder: validSortOrder,
+        modelFilter
       })
     }
 
@@ -322,7 +377,17 @@ router.get('/api-keys', authenticateAdmin, async (req, res) => {
  * 使用预计算索引进行费用排序的分页查询
  */
 async function getApiKeysSortedByCostPrecomputed(options) {
-  const { page, pageSize, sortOrder, costTimeRange, search, searchMode, tag, isActive } = options
+  const {
+    page,
+    pageSize,
+    sortOrder,
+    costTimeRange,
+    search,
+    searchMode,
+    tag,
+    isActive,
+    modelFilter = []
+  } = options
   const costRankService = require('../../services/costRankService')
 
   // 1. 获取排序后的全量 keyId 列表
@@ -369,6 +434,15 @@ async function getApiKeysSortedByCostPrecomputed(options) {
     }
   }
 
+  // 模型筛选
+  if (modelFilter.length > 0) {
+    const keyIdsWithModels = await redis.getKeyIdsWithModels(
+      orderedKeys.map((k) => k.id),
+      modelFilter
+    )
+    orderedKeys = orderedKeys.filter((k) => keyIdsWithModels.has(k.id))
+  }
+
   // 5. 收集所有可用标签
   const allTags = new Set()
   for (const key of allKeys) {
@@ -411,8 +485,18 @@ async function getApiKeysSortedByCostPrecomputed(options) {
  * 使用实时计算进行 custom 时间范围的费用排序
  */
 async function getApiKeysSortedByCostCustom(options) {
-  const { page, pageSize, sortOrder, startDate, endDate, search, searchMode, tag, isActive } =
-    options
+  const {
+    page,
+    pageSize,
+    sortOrder,
+    startDate,
+    endDate,
+    search,
+    searchMode,
+    tag,
+    isActive,
+    modelFilter = []
+  } = options
   const costRankService = require('../../services/costRankService')
 
   // 1. 实时计算所有 Keys 的费用
@@ -427,9 +511,9 @@ async function getApiKeysSortedByCostCustom(options) {
   }
 
   // 2. 转换为数组并排序
-  const sortedEntries = [...costs.entries()].sort((a, b) => {
-    return sortOrder === 'desc' ? b[1] - a[1] : a[1] - b[1]
-  })
+  const sortedEntries = [...costs.entries()].sort((a, b) =>
+    sortOrder === 'desc' ? b[1] - a[1] : a[1] - b[1]
+  )
   const rankedKeyIds = sortedEntries.map(([keyId]) => keyId)
 
   // 3. 批量获取 API Key 基础数据
@@ -465,6 +549,15 @@ async function getApiKeysSortedByCostCustom(options) {
     }
   }
 
+  // 模型筛选
+  if (modelFilter.length > 0) {
+    const keyIdsWithModels = await redis.getKeyIdsWithModels(
+      orderedKeys.map((k) => k.id),
+      modelFilter
+    )
+    orderedKeys = orderedKeys.filter((k) => keyIdsWithModels.has(k.id))
+  }
+
   // 6. 收集所有可用标签
   const allTags = new Set()
   for (const key of allKeys) {
@@ -863,6 +956,86 @@ async function calculateKeyStats(keyId, timeRange, startDate, endDate) {
   // 去重（避免日数据和月数据重复计算）
   const uniqueKeys = [...new Set(allKeys)]
 
+  // 获取实时限制数据（窗口数据不受时间范围筛选影响，始终获取当前窗口状态）
+  let dailyCost = 0
+  let currentWindowCost = 0
+  let windowRemainingSeconds = null
+  let windowStartTime = null
+  let windowEndTime = null
+  let allTimeCost = 0
+
+  try {
+    // 先获取 API Key 配置，判断是否需要查询限制相关数据
+    const apiKey = await redis.getApiKey(keyId)
+    const rateLimitWindow = parseInt(apiKey?.rateLimitWindow) || 0
+    const dailyCostLimit = parseFloat(apiKey?.dailyCostLimit) || 0
+    const totalCostLimit = parseFloat(apiKey?.totalCostLimit) || 0
+
+    // 只在启用了每日费用限制时查询
+    if (dailyCostLimit > 0) {
+      dailyCost = await redis.getDailyCost(keyId)
+    }
+
+    // 只在启用了总费用限制时查询
+    if (totalCostLimit > 0) {
+      const totalCostKey = `usage:cost:total:${keyId}`
+      allTimeCost = parseFloat((await client.get(totalCostKey)) || '0')
+    }
+
+    // 🔧 FIX: 对于 "全部时间" 时间范围，直接使用 allTimeCost
+    // 因为 usage:*:model:daily:* 键有 30 天 TTL，旧数据已经过期
+    if (timeRange === 'all' && allTimeCost > 0) {
+      logger.debug(`📊 使用 allTimeCost 计算 timeRange='all': ${allTimeCost}`)
+
+      return {
+        requests: 0, // 旧数据详情不可用
+        tokens: 0,
+        inputTokens: 0,
+        outputTokens: 0,
+        cacheCreateTokens: 0,
+        cacheReadTokens: 0,
+        cost: allTimeCost,
+        formattedCost: CostCalculator.formatCost(allTimeCost),
+        // 实时限制数据（始终返回，不受时间范围影响）
+        dailyCost,
+        currentWindowCost,
+        windowRemainingSeconds,
+        windowStartTime,
+        windowEndTime,
+        allTimeCost
+      }
+    }
+
+    // 只在启用了窗口限制时查询窗口数据
+    if (rateLimitWindow > 0) {
+      const costCountKey = `rate_limit:cost:${keyId}`
+      const windowStartKey = `rate_limit:window_start:${keyId}`
+
+      currentWindowCost = parseFloat((await client.get(costCountKey)) || '0')
+
+      // 获取窗口开始时间和计算剩余时间
+      const windowStart = await client.get(windowStartKey)
+      if (windowStart) {
+        const now = Date.now()
+        windowStartTime = parseInt(windowStart)
+        const windowDuration = rateLimitWindow * 60 * 1000 // 转换为毫秒
+        windowEndTime = windowStartTime + windowDuration
+
+        // 如果窗口还有效
+        if (now < windowEndTime) {
+          windowRemainingSeconds = Math.max(0, Math.floor((windowEndTime - now) / 1000))
+        } else {
+          // 窗口已过期
+          windowRemainingSeconds = 0
+          currentWindowCost = 0
+        }
+      }
+    }
+  } catch (error) {
+    logger.warn(`⚠️ 获取实时限制数据失败 (key: ${keyId}):`, error.message)
+  }
+
+  // 如果没有使用数据，返回零值但包含窗口数据
   if (uniqueKeys.length === 0) {
     return {
       requests: 0,
@@ -872,7 +1045,14 @@ async function calculateKeyStats(keyId, timeRange, startDate, endDate) {
       cacheCreateTokens: 0,
       cacheReadTokens: 0,
       cost: 0,
-      formattedCost: '$0.00'
+      formattedCost: '$0.00',
+      // 实时限制数据（始终返回，不受时间范围影响）
+      dailyCost,
+      currentWindowCost,
+      windowRemainingSeconds,
+      windowStartTime,
+      windowEndTime,
+      allTimeCost
     }
   }
 
@@ -887,12 +1067,10 @@ async function calculateKeyStats(keyId, timeRange, startDate, endDate) {
   const modelStatsMap = new Map()
   let totalRequests = 0
 
-  // 用于去重：只统计日数据，避免与月数据重复
+  // 用于去重：先统计月数据，避免与日数据重复
   const dailyKeyPattern = /usage:.+:model:daily:(.+):\d{4}-\d{2}-\d{2}$/
   const monthlyKeyPattern = /usage:.+:model:monthly:(.+):\d{4}-\d{2}$/
-
-  // 检查是否有日数据
-  const hasDailyData = uniqueKeys.some((key) => dailyKeyPattern.test(key))
+  const currentMonth = `${tzDate.getUTCFullYear()}-${String(tzDate.getUTCMonth() + 1).padStart(2, '0')}`
 
   for (let i = 0; i < results.length; i++) {
     const [err, data] = results[i]
@@ -919,8 +1097,12 @@ async function calculateKeyStats(keyId, timeRange, startDate, endDate) {
       continue
     }
 
-    // 如果有日数据，则跳过月数据以避免重复
-    if (hasDailyData && isMonthly) {
+    // 跳过当前月的月数据
+    if (isMonthly && key.includes(`:${currentMonth}`)) {
+      continue
+    }
+    // 跳过非当前月的日数据
+    if (!isMonthly && !key.includes(`:${currentMonth}-`)) {
       continue
     }
 
@@ -973,52 +1155,6 @@ async function calculateKeyStats(keyId, timeRange, startDate, endDate) {
 
   const tokens = inputTokens + outputTokens + cacheCreateTokens + cacheReadTokens
 
-  // 获取实时限制数据
-  let dailyCost = 0
-  let currentWindowCost = 0
-  let windowRemainingSeconds = null
-  let windowStartTime = null
-  let windowEndTime = null
-  let allTimeCost = 0
-
-  try {
-    // 获取当日费用
-    dailyCost = await redis.getDailyCost(keyId)
-
-    // 获取历史总费用（用于总费用限制进度条，不受时间范围影响）
-    const totalCostKey = `usage:cost:total:${keyId}`
-    allTimeCost = parseFloat((await client.get(totalCostKey)) || '0')
-
-    // 获取 API Key 配置信息以判断是否需要窗口数据
-    const apiKey = await redis.getApiKey(keyId)
-    if (apiKey && apiKey.rateLimitWindow > 0) {
-      const costCountKey = `rate_limit:cost:${keyId}`
-      const windowStartKey = `rate_limit:window_start:${keyId}`
-
-      currentWindowCost = parseFloat((await client.get(costCountKey)) || '0')
-
-      // 获取窗口开始时间和计算剩余时间
-      const windowStart = await client.get(windowStartKey)
-      if (windowStart) {
-        const now = Date.now()
-        windowStartTime = parseInt(windowStart)
-        const windowDuration = apiKey.rateLimitWindow * 60 * 1000 // 转换为毫秒
-        windowEndTime = windowStartTime + windowDuration
-
-        // 如果窗口还有效
-        if (now < windowEndTime) {
-          windowRemainingSeconds = Math.max(0, Math.floor((windowEndTime - now) / 1000))
-        } else {
-          // 窗口已过期
-          windowRemainingSeconds = 0
-          currentWindowCost = 0
-        }
-      }
-    }
-  } catch (error) {
-    logger.debug(`获取实时限制数据失败 (key: ${keyId}):`, error.message)
-  }
-
   return {
     requests: totalRequests,
     tokens,
@@ -1283,16 +1419,10 @@ router.post('/api-keys', authenticateAdmin, async (req, res) => {
       }
     }
 
-    // 验证服务权限字段
-    if (
-      permissions !== undefined &&
-      permissions !== null &&
-      permissions !== '' &&
-      !['claude', 'gemini', 'openai', 'droid', 'all'].includes(permissions)
-    ) {
-      return res.status(400).json({
-        error: 'Invalid permissions value. Must be claude, gemini, openai, droid, or all'
-      })
+    // 验证服务权限字段（支持数组格式）
+    const permissionsError = validatePermissions(permissions)
+    if (permissionsError) {
+      return res.status(400).json({ error: permissionsError })
     }
 
     const newKey = await apiKeyService.generateApiKey({
@@ -1382,15 +1512,10 @@ router.post('/api-keys/batch', authenticateAdmin, async (req, res) => {
         .json({ error: 'Base name must be less than 90 characters to allow for numbering' })
     }
 
-    if (
-      permissions !== undefined &&
-      permissions !== null &&
-      permissions !== '' &&
-      !['claude', 'gemini', 'openai', 'droid', 'all'].includes(permissions)
-    ) {
-      return res.status(400).json({
-        error: 'Invalid permissions value. Must be claude, gemini, openai, droid, or all'
-      })
+    // 验证服务权限字段（支持数组格式）
+    const batchPermissionsError = validatePermissions(permissions)
+    if (batchPermissionsError) {
+      return res.status(400).json({ error: batchPermissionsError })
     }
 
     // 生成批量API Keys
@@ -1493,13 +1618,12 @@ router.put('/api-keys/batch', authenticateAdmin, async (req, res) => {
       })
     }
 
-    if (
-      updates.permissions !== undefined &&
-      !['claude', 'gemini', 'openai', 'droid', 'all'].includes(updates.permissions)
-    ) {
-      return res.status(400).json({
-        error: 'Invalid permissions value. Must be claude, gemini, openai, droid, or all'
-      })
+    // 验证服务权限字段（支持数组格式）
+    if (updates.permissions !== undefined) {
+      const updatePermissionsError = validatePermissions(updates.permissions)
+      if (updatePermissionsError) {
+        return res.status(400).json({ error: updatePermissionsError })
+      }
     }
 
     logger.info(
@@ -1774,11 +1898,10 @@ router.put('/api-keys/:keyId', authenticateAdmin, async (req, res) => {
     }
 
     if (permissions !== undefined) {
-      // 验证权限值
-      if (!['claude', 'gemini', 'openai', 'droid', 'all'].includes(permissions)) {
-        return res.status(400).json({
-          error: 'Invalid permissions value. Must be claude, gemini, openai, droid, or all'
-        })
+      // 验证服务权限字段（支持数组格式）
+      const singlePermissionsError = validatePermissions(permissions)
+      if (singlePermissionsError) {
+        return res.status(400).json({ error: singlePermissionsError })
       }
       updates.permissions = permissions
     }

src/routes/admin/balanceScripts.js

@@ -0,0 +1,41 @@
+const express = require('express')
+const { authenticateAdmin } = require('../../middleware/auth')
+const balanceScriptService = require('../../services/balanceScriptService')
+const router = express.Router()
+
+// 获取全部脚本配置列表
+router.get('/balance-scripts', authenticateAdmin, (req, res) => {
+  const items = balanceScriptService.listConfigs()
+  return res.json({ success: true, data: items })
+})
+
+// 获取单个脚本配置
+router.get('/balance-scripts/:name', authenticateAdmin, (req, res) => {
+  const { name } = req.params
+  const config = balanceScriptService.getConfig(name || 'default')
+  return res.json({ success: true, data: config })
+})
+
+// 保存脚本配置
+router.put('/balance-scripts/:name', authenticateAdmin, (req, res) => {
+  try {
+    const { name } = req.params
+    const saved = balanceScriptService.saveConfig(name || 'default', req.body || {})
+    return res.json({ success: true, data: saved })
+  } catch (error) {
+    return res.status(400).json({ success: false, error: error.message })
+  }
+})
+
+// 测试脚本（不落库）
+router.post('/balance-scripts/:name/test', authenticateAdmin, async (req, res) => {
+  try {
+    const { name } = req.params
+    const result = await balanceScriptService.testScript(name || 'default', req.body || {})
+    return res.json({ success: true, data: result })
+  } catch (error) {
+    return res.status(400).json({ success: false, error: error.message })
+  }
+})
+
+module.exports = router

src/routes/admin/bedrockAccounts.js

@@ -122,6 +122,7 @@ router.post('/', authenticateAdmin, async (req, res) => {
       description,
       region,
       awsCredentials,
+      bearerToken,
       defaultModel,
       priority,
       accountType,
@@ -145,9 +146,9 @@ router.post('/', authenticateAdmin, async (req, res) => {
     }
 
     // 验证credentialType的有效性
-    if (credentialType && !['default', 'access_key', 'bearer_token'].includes(credentialType)) {
+    if (credentialType && !['access_key', 'bearer_token'].includes(credentialType)) {
       return res.status(400).json({
-        error: 'Invalid credential type. Must be "default", "access_key", or "bearer_token"'
+        error: 'Invalid credential type. Must be "access_key" or "bearer_token"'
       })
     }
 
@@ -156,10 +157,11 @@ router.post('/', authenticateAdmin, async (req, res) => {
       description: description || '',
       region: region || 'us-east-1',
       awsCredentials,
+      bearerToken,
       defaultModel,
       priority: priority || 50,
       accountType: accountType || 'shared',
-      credentialType: credentialType || 'default'
+      credentialType: credentialType || 'access_key'
     })
 
     if (!result.success) {
@@ -206,10 +208,10 @@ router.put('/:accountId', authenticateAdmin, async (req, res) => {
     // 验证credentialType的有效性
     if (
       mappedUpdates.credentialType &&
-      !['default', 'access_key', 'bearer_token'].includes(mappedUpdates.credentialType)
+      !['access_key', 'bearer_token'].includes(mappedUpdates.credentialType)
     ) {
       return res.status(400).json({
-        error: 'Invalid credential type. Must be "default", "access_key", or "bearer_token"'
+        error: 'Invalid credential type. Must be "access_key" or "bearer_token"'
       })
     }
 
@@ -349,22 +351,15 @@ router.put('/:accountId/toggle-schedulable', authenticateAdmin, async (req, res)
   }
 })
 
-// 测试Bedrock账户连接
+// 测试Bedrock账户连接（SSE 流式）
 router.post('/:accountId/test', authenticateAdmin, async (req, res) => {
   try {
     const { accountId } = req.params
 
-    const result = await bedrockAccountService.testAccount(accountId)
-
-    if (!result.success) {
-      return res.status(500).json({ error: 'Account test failed', message: result.error })
-    }
-
-    logger.success(`🧪 Admin tested Bedrock account: ${accountId} - ${result.data.status}`)
-    return res.json({ success: true, data: result.data })
+    await bedrockAccountService.testAccountConnection(accountId, res)
   } catch (error) {
     logger.error('❌ Failed to test Bedrock account:', error)
-    return res.status(500).json({ error: 'Failed to test Bedrock account', message: error.message })
+    // 错误已在服务层处理，这里仅做日志记录
   }
 })
 

src/routes/admin/claudeAccounts.js

@@ -9,6 +9,7 @@ const router = express.Router()
 const claudeAccountService = require('../../services/claudeAccountService')
 const claudeRelayService = require('../../services/claudeRelayService')
 const accountGroupService = require('../../services/accountGroupService')
+const accountTestSchedulerService = require('../../services/accountTestSchedulerService')
 const apiKeyService = require('../../services/apiKeyService')
 const redis = require('../../models/redis')
 const { authenticateAdmin } = require('../../middleware/auth')
@@ -255,6 +256,108 @@ router.post('/claude-accounts/exchange-setup-token-code', authenticateAdmin, asy
   }
 })
 
+// =============================================================================
+// Cookie自动授权端点 (基于sessionKey自动完成OAuth流程)
+// =============================================================================
+
+// 普通OAuth的Cookie自动授权
+router.post('/claude-accounts/oauth-with-cookie', authenticateAdmin, async (req, res) => {
+  try {
+    const { sessionKey, proxy } = req.body
+
+    // 验证sessionKey参数
+    if (!sessionKey || typeof sessionKey !== 'string' || sessionKey.trim().length === 0) {
+      return res.status(400).json({
+        success: false,
+        error: 'sessionKey不能为空',
+        message: '请提供有效的sessionKey值'
+      })
+    }
+
+    const trimmedSessionKey = sessionKey.trim()
+
+    logger.info('🍪 Starting Cookie-based OAuth authorization', {
+      sessionKeyLength: trimmedSessionKey.length,
+      sessionKeyPrefix: `${trimmedSessionKey.substring(0, 10)}...`,
+      hasProxy: !!proxy
+    })
+
+    // 执行Cookie自动授权流程
+    const result = await oauthHelper.oauthWithCookie(trimmedSessionKey, proxy, false)
+
+    logger.success('🎉 Cookie-based OAuth authorization completed successfully')
+
+    return res.json({
+      success: true,
+      data: {
+        claudeAiOauth: result.claudeAiOauth,
+        organizationUuid: result.organizationUuid,
+        capabilities: result.capabilities
+      }
+    })
+  } catch (error) {
+    logger.error('❌ Cookie-based OAuth authorization failed:', {
+      error: error.message,
+      sessionKeyLength: req.body.sessionKey ? req.body.sessionKey.length : 0
+    })
+
+    return res.status(500).json({
+      success: false,
+      error: 'Cookie授权失败',
+      message: error.message
+    })
+  }
+})
+
+// Setup Token的Cookie自动授权
+router.post('/claude-accounts/setup-token-with-cookie', authenticateAdmin, async (req, res) => {
+  try {
+    const { sessionKey, proxy } = req.body
+
+    // 验证sessionKey参数
+    if (!sessionKey || typeof sessionKey !== 'string' || sessionKey.trim().length === 0) {
+      return res.status(400).json({
+        success: false,
+        error: 'sessionKey不能为空',
+        message: '请提供有效的sessionKey值'
+      })
+    }
+
+    const trimmedSessionKey = sessionKey.trim()
+
+    logger.info('🍪 Starting Cookie-based Setup Token authorization', {
+      sessionKeyLength: trimmedSessionKey.length,
+      sessionKeyPrefix: `${trimmedSessionKey.substring(0, 10)}...`,
+      hasProxy: !!proxy
+    })
+
+    // 执行Cookie自动授权流程（Setup Token模式）
+    const result = await oauthHelper.oauthWithCookie(trimmedSessionKey, proxy, true)
+
+    logger.success('🎉 Cookie-based Setup Token authorization completed successfully')
+
+    return res.json({
+      success: true,
+      data: {
+        claudeAiOauth: result.claudeAiOauth,
+        organizationUuid: result.organizationUuid,
+        capabilities: result.capabilities
+      }
+    })
+  } catch (error) {
+    logger.error('❌ Cookie-based Setup Token authorization failed:', {
+      error: error.message,
+      sessionKeyLength: req.body.sessionKey ? req.body.sessionKey.length : 0
+    })
+
+    return res.status(500).json({
+      success: false,
+      error: 'Cookie授权失败',
+      message: error.message
+    })
+  }
+})
+
 // 获取所有Claude账户
 router.get('/claude-accounts', authenticateAdmin, async (req, res) => {
   try {
@@ -481,7 +584,9 @@ router.post('/claude-accounts', authenticateAdmin, async (req, res) => {
       useUnifiedClientId,
       unifiedClientId,
       expiresAt,
-      extInfo
+      extInfo,
+      maxConcurrency,
+      interceptWarmup
     } = req.body
 
     if (!name) {
@@ -526,7 +631,9 @@ router.post('/claude-accounts', authenticateAdmin, async (req, res) => {
       useUnifiedClientId: useUnifiedClientId === true, // 默认为false
       unifiedClientId: unifiedClientId || '', // 统一的客户端标识
       expiresAt: expiresAt || null, // 账户订阅到期时间
-      extInfo: extInfo || null
+      extInfo: extInfo || null,
+      maxConcurrency: maxConcurrency || 0, // 账户级串行队列：0=使用全局配置，>0=强制启用
+      interceptWarmup: interceptWarmup === true // 拦截预热请求：默认为false
     })
 
     // 如果是分组类型，将账户添加到分组
@@ -801,4 +908,219 @@ router.post('/claude-accounts/:accountId/test', authenticateAdmin, async (req, r
   }
 })
 
+// ============================================================================
+// 账户定时测试相关端点
+// ============================================================================
+
+// 获取账户测试历史
+router.get('/claude-accounts/:accountId/test-history', authenticateAdmin, async (req, res) => {
+  const { accountId } = req.params
+
+  try {
+    const history = await redis.getAccountTestHistory(accountId, 'claude')
+    return res.json({
+      success: true,
+      data: {
+        accountId,
+        platform: 'claude',
+        history
+      }
+    })
+  } catch (error) {
+    logger.error(`❌ Failed to get test history for account ${accountId}:`, error)
+    return res.status(500).json({
+      error: 'Failed to get test history',
+      message: error.message
+    })
+  }
+})
+
+// 获取账户定时测试配置
+router.get('/claude-accounts/:accountId/test-config', authenticateAdmin, async (req, res) => {
+  const { accountId } = req.params
+
+  try {
+    const testConfig = await redis.getAccountTestConfig(accountId, 'claude')
+    return res.json({
+      success: true,
+      data: {
+        accountId,
+        platform: 'claude',
+        config: testConfig || {
+          enabled: false,
+          cronExpression: '0 8 * * *',
+          model: 'claude-sonnet-4-5-20250929'
+        }
+      }
+    })
+  } catch (error) {
+    logger.error(`❌ Failed to get test config for account ${accountId}:`, error)
+    return res.status(500).json({
+      error: 'Failed to get test config',
+      message: error.message
+    })
+  }
+})
+
+// 设置账户定时测试配置
+router.put('/claude-accounts/:accountId/test-config', authenticateAdmin, async (req, res) => {
+  const { accountId } = req.params
+  const { enabled, cronExpression, model } = req.body
+
+  try {
+    // 验证 enabled 参数
+    if (typeof enabled !== 'boolean') {
+      return res.status(400).json({
+        error: 'Invalid parameter',
+        message: 'enabled must be a boolean'
+      })
+    }
+
+    // 验证 cronExpression 参数
+    if (!cronExpression || typeof cronExpression !== 'string') {
+      return res.status(400).json({
+        error: 'Invalid parameter',
+        message: 'cronExpression is required and must be a string'
+      })
+    }
+
+    // 限制 cronExpression 长度防止 DoS
+    const MAX_CRON_LENGTH = 100
+    if (cronExpression.length > MAX_CRON_LENGTH) {
+      return res.status(400).json({
+        error: 'Invalid parameter',
+        message: `cronExpression too long (max ${MAX_CRON_LENGTH} characters)`
+      })
+    }
+
+    // 使用 service 的方法验证 cron 表达式
+    if (!accountTestSchedulerService.validateCronExpression(cronExpression)) {
+      return res.status(400).json({
+        error: 'Invalid parameter',
+        message: `Invalid cron expression: ${cronExpression}. Format: "minute hour day month weekday" (e.g., "0 8 * * *" for daily at 8:00)`
+      })
+    }
+
+    // 验证模型参数
+    const testModel = model || 'claude-sonnet-4-5-20250929'
+    if (typeof testModel !== 'string' || testModel.length > 256) {
+      return res.status(400).json({
+        error: 'Invalid parameter',
+        message: 'model must be a valid string (max 256 characters)'
+      })
+    }
+
+    // 检查账户是否存在
+    const account = await claudeAccountService.getAccount(accountId)
+    if (!account) {
+      return res.status(404).json({
+        error: 'Account not found',
+        message: `Claude account ${accountId} not found`
+      })
+    }
+
+    // 保存配置
+    await redis.saveAccountTestConfig(accountId, 'claude', {
+      enabled,
+      cronExpression,
+      model: testModel
+    })
+
+    logger.success(
+      `📝 Updated test config for Claude account ${accountId}: enabled=${enabled}, cronExpression=${cronExpression}, model=${testModel}`
+    )
+
+    return res.json({
+      success: true,
+      message: 'Test config updated successfully',
+      data: {
+        accountId,
+        platform: 'claude',
+        config: { enabled, cronExpression, model: testModel }
+      }
+    })
+  } catch (error) {
+    logger.error(`❌ Failed to update test config for account ${accountId}:`, error)
+    return res.status(500).json({
+      error: 'Failed to update test config',
+      message: error.message
+    })
+  }
+})
+
+// 手动触发账户测试（非流式，返回JSON结果）
+router.post('/claude-accounts/:accountId/test-sync', authenticateAdmin, async (req, res) => {
+  const { accountId } = req.params
+
+  try {
+    // 检查账户是否存在
+    const account = await claudeAccountService.getAccount(accountId)
+    if (!account) {
+      return res.status(404).json({
+        error: 'Account not found',
+        message: `Claude account ${accountId} not found`
+      })
+    }
+
+    logger.info(`🧪 Manual sync test triggered for Claude account: ${accountId}`)
+
+    // 执行测试
+    const testResult = await claudeRelayService.testAccountConnectionSync(accountId)
+
+    // 保存测试结果到历史
+    await redis.saveAccountTestResult(accountId, 'claude', testResult)
+    await redis.setAccountLastTestTime(accountId, 'claude')
+
+    return res.json({
+      success: true,
+      data: {
+        accountId,
+        platform: 'claude',
+        result: testResult
+      }
+    })
+  } catch (error) {
+    logger.error(`❌ Failed to run sync test for account ${accountId}:`, error)
+    return res.status(500).json({
+      error: 'Failed to run test',
+      message: error.message
+    })
+  }
+})
+
+// 批量获取多个账户的测试历史
+router.post('/claude-accounts/batch-test-history', authenticateAdmin, async (req, res) => {
+  const { accountIds } = req.body
+
+  try {
+    if (!Array.isArray(accountIds) || accountIds.length === 0) {
+      return res.status(400).json({
+        error: 'Invalid parameter',
+        message: 'accountIds must be a non-empty array'
+      })
+    }
+
+    // 限制批量查询数量
+    const limitedIds = accountIds.slice(0, 100)
+
+    const accounts = limitedIds.map((accountId) => ({
+      accountId,
+      platform: 'claude'
+    }))
+
+    const historyMap = await redis.getAccountsTestHistory(accounts)
+
+    return res.json({
+      success: true,
+      data: historyMap
+    })
+  } catch (error) {
+    logger.error('❌ Failed to get batch test history:', error)
+    return res.status(500).json({
+      error: 'Failed to get batch test history',
+      message: error.message
+    })
+  }
+})
+
 module.exports = router

src/routes/admin/claudeConsoleAccounts.js

@@ -131,7 +131,9 @@ router.post('/claude-console-accounts', authenticateAdmin, async (req, res) => {
       groupId,
       dailyQuota,
       quotaResetTime,
-      maxConcurrentTasks
+      maxConcurrentTasks,
+      disableAutoProtection,
+      interceptWarmup
     } = req.body
 
     if (!name || !apiUrl || !apiKey) {
@@ -151,6 +153,10 @@ router.post('/claude-console-accounts', authenticateAdmin, async (req, res) => {
       }
     }
 
+    // 校验上游错误自动防护开关
+    const normalizedDisableAutoProtection =
+      disableAutoProtection === true || disableAutoProtection === 'true'
+
     // 验证accountType的有效性
     if (accountType && !['shared', 'dedicated', 'group'].includes(accountType)) {
       return res
@@ -180,7 +186,9 @@ router.post('/claude-console-accounts', authenticateAdmin, async (req, res) => {
       maxConcurrentTasks:
         maxConcurrentTasks !== undefined && maxConcurrentTasks !== null
           ? Number(maxConcurrentTasks)
-          : 0
+          : 0,
+      disableAutoProtection: normalizedDisableAutoProtection,
+      interceptWarmup: interceptWarmup === true || interceptWarmup === 'true'
     })
 
     // 如果是分组类型，将账户添加到分组（CCR 归属 Claude 平台分组）
@@ -250,6 +258,13 @@ router.put('/claude-console-accounts/:accountId', authenticateAdmin, async (req,
       return res.status(404).json({ error: 'Account not found' })
     }
 
+    // 规范化上游错误自动防护开关
+    if (mappedUpdates.disableAutoProtection !== undefined) {
+      mappedUpdates.disableAutoProtection =
+        mappedUpdates.disableAutoProtection === true ||
+        mappedUpdates.disableAutoProtection === 'true'
+    }
+
     // 处理分组的变更
     if (mappedUpdates.accountType !== undefined) {
       // 如果之前是分组类型，需要从所有分组中移除

src/routes/admin/claudeRelayConfig.js

@@ -0,0 +1,239 @@
+/**
+ * Claude 转发配置 API 路由
+ * 管理全局 Claude Code 限制和会话绑定配置
+ */
+
+const express = require('express')
+const { authenticateAdmin } = require('../../middleware/auth')
+const claudeRelayConfigService = require('../../services/claudeRelayConfigService')
+const logger = require('../../utils/logger')
+
+const router = express.Router()
+
+/**
+ * GET /admin/claude-relay-config
+ * 获取 Claude 转发配置
+ */
+router.get('/claude-relay-config', authenticateAdmin, async (req, res) => {
+  try {
+    const config = await claudeRelayConfigService.getConfig()
+    return res.json({
+      success: true,
+      config
+    })
+  } catch (error) {
+    logger.error('❌ Failed to get Claude relay config:', error)
+    return res.status(500).json({
+      error: 'Failed to get configuration',
+      message: error.message
+    })
+  }
+})
+
+/**
+ * PUT /admin/claude-relay-config
+ * 更新 Claude 转发配置
+ */
+router.put('/claude-relay-config', authenticateAdmin, async (req, res) => {
+  try {
+    const {
+      claudeCodeOnlyEnabled,
+      globalSessionBindingEnabled,
+      sessionBindingErrorMessage,
+      sessionBindingTtlDays,
+      userMessageQueueEnabled,
+      userMessageQueueDelayMs,
+      userMessageQueueTimeoutMs,
+      concurrentRequestQueueEnabled,
+      concurrentRequestQueueMaxSize,
+      concurrentRequestQueueMaxSizeMultiplier,
+      concurrentRequestQueueTimeoutMs
+    } = req.body
+
+    // 验证输入
+    if (claudeCodeOnlyEnabled !== undefined && typeof claudeCodeOnlyEnabled !== 'boolean') {
+      return res.status(400).json({ error: 'claudeCodeOnlyEnabled must be a boolean' })
+    }
+
+    if (
+      globalSessionBindingEnabled !== undefined &&
+      typeof globalSessionBindingEnabled !== 'boolean'
+    ) {
+      return res.status(400).json({ error: 'globalSessionBindingEnabled must be a boolean' })
+    }
+
+    if (sessionBindingErrorMessage !== undefined) {
+      if (typeof sessionBindingErrorMessage !== 'string') {
+        return res.status(400).json({ error: 'sessionBindingErrorMessage must be a string' })
+      }
+      if (sessionBindingErrorMessage.length > 500) {
+        return res
+          .status(400)
+          .json({ error: 'sessionBindingErrorMessage must be less than 500 characters' })
+      }
+    }
+
+    if (sessionBindingTtlDays !== undefined) {
+      if (
+        typeof sessionBindingTtlDays !== 'number' ||
+        sessionBindingTtlDays < 1 ||
+        sessionBindingTtlDays > 365
+      ) {
+        return res
+          .status(400)
+          .json({ error: 'sessionBindingTtlDays must be a number between 1 and 365' })
+      }
+    }
+
+    // 验证用户消息队列配置
+    if (userMessageQueueEnabled !== undefined && typeof userMessageQueueEnabled !== 'boolean') {
+      return res.status(400).json({ error: 'userMessageQueueEnabled must be a boolean' })
+    }
+
+    if (userMessageQueueDelayMs !== undefined) {
+      if (
+        typeof userMessageQueueDelayMs !== 'number' ||
+        userMessageQueueDelayMs < 0 ||
+        userMessageQueueDelayMs > 10000
+      ) {
+        return res
+          .status(400)
+          .json({ error: 'userMessageQueueDelayMs must be a number between 0 and 10000' })
+      }
+    }
+
+    if (userMessageQueueTimeoutMs !== undefined) {
+      if (
+        typeof userMessageQueueTimeoutMs !== 'number' ||
+        userMessageQueueTimeoutMs < 1000 ||
+        userMessageQueueTimeoutMs > 300000
+      ) {
+        return res
+          .status(400)
+          .json({ error: 'userMessageQueueTimeoutMs must be a number between 1000 and 300000' })
+      }
+    }
+
+    // 验证并发请求排队配置
+    if (
+      concurrentRequestQueueEnabled !== undefined &&
+      typeof concurrentRequestQueueEnabled !== 'boolean'
+    ) {
+      return res.status(400).json({ error: 'concurrentRequestQueueEnabled must be a boolean' })
+    }
+
+    if (concurrentRequestQueueMaxSize !== undefined) {
+      if (
+        typeof concurrentRequestQueueMaxSize !== 'number' ||
+        !Number.isInteger(concurrentRequestQueueMaxSize) ||
+        concurrentRequestQueueMaxSize < 1 ||
+        concurrentRequestQueueMaxSize > 100
+      ) {
+        return res
+          .status(400)
+          .json({ error: 'concurrentRequestQueueMaxSize must be an integer between 1 and 100' })
+      }
+    }
+
+    if (concurrentRequestQueueMaxSizeMultiplier !== undefined) {
+      // 使用 Number.isFinite() 同时排除 NaN、Infinity、-Infinity 和非数字类型
+      if (
+        !Number.isFinite(concurrentRequestQueueMaxSizeMultiplier) ||
+        concurrentRequestQueueMaxSizeMultiplier < 0 ||
+        concurrentRequestQueueMaxSizeMultiplier > 10
+      ) {
+        return res.status(400).json({
+          error: 'concurrentRequestQueueMaxSizeMultiplier must be a finite number between 0 and 10'
+        })
+      }
+    }
+
+    if (concurrentRequestQueueTimeoutMs !== undefined) {
+      if (
+        typeof concurrentRequestQueueTimeoutMs !== 'number' ||
+        !Number.isInteger(concurrentRequestQueueTimeoutMs) ||
+        concurrentRequestQueueTimeoutMs < 5000 ||
+        concurrentRequestQueueTimeoutMs > 300000
+      ) {
+        return res.status(400).json({
+          error:
+            'concurrentRequestQueueTimeoutMs must be an integer between 5000 and 300000 (5 seconds to 5 minutes)'
+        })
+      }
+    }
+
+    const updateData = {}
+    if (claudeCodeOnlyEnabled !== undefined) {
+      updateData.claudeCodeOnlyEnabled = claudeCodeOnlyEnabled
+    }
+    if (globalSessionBindingEnabled !== undefined) {
+      updateData.globalSessionBindingEnabled = globalSessionBindingEnabled
+    }
+    if (sessionBindingErrorMessage !== undefined) {
+      updateData.sessionBindingErrorMessage = sessionBindingErrorMessage
+    }
+    if (sessionBindingTtlDays !== undefined) {
+      updateData.sessionBindingTtlDays = sessionBindingTtlDays
+    }
+    if (userMessageQueueEnabled !== undefined) {
+      updateData.userMessageQueueEnabled = userMessageQueueEnabled
+    }
+    if (userMessageQueueDelayMs !== undefined) {
+      updateData.userMessageQueueDelayMs = userMessageQueueDelayMs
+    }
+    if (userMessageQueueTimeoutMs !== undefined) {
+      updateData.userMessageQueueTimeoutMs = userMessageQueueTimeoutMs
+    }
+    if (concurrentRequestQueueEnabled !== undefined) {
+      updateData.concurrentRequestQueueEnabled = concurrentRequestQueueEnabled
+    }
+    if (concurrentRequestQueueMaxSize !== undefined) {
+      updateData.concurrentRequestQueueMaxSize = concurrentRequestQueueMaxSize
+    }
+    if (concurrentRequestQueueMaxSizeMultiplier !== undefined) {
+      updateData.concurrentRequestQueueMaxSizeMultiplier = concurrentRequestQueueMaxSizeMultiplier
+    }
+    if (concurrentRequestQueueTimeoutMs !== undefined) {
+      updateData.concurrentRequestQueueTimeoutMs = concurrentRequestQueueTimeoutMs
+    }
+
+    const updatedConfig = await claudeRelayConfigService.updateConfig(
+      updateData,
+      req.admin?.username || 'unknown'
+    )
+
+    return res.json({
+      success: true,
+      message: 'Configuration updated successfully',
+      config: updatedConfig
+    })
+  } catch (error) {
+    logger.error('❌ Failed to update Claude relay config:', error)
+    return res.status(500).json({
+      error: 'Failed to update configuration',
+      message: error.message
+    })
+  }
+})
+
+/**
+ * GET /admin/claude-relay-config/session-bindings
+ * 获取会话绑定统计
+ */
+router.get('/claude-relay-config/session-bindings', authenticateAdmin, async (req, res) => {
+  try {
+    const stats = await claudeRelayConfigService.getSessionBindingStats()
+    return res.json({
+      success: true,
+      data: stats
+    })
+  } catch (error) {
+    logger.error('❌ Failed to get session binding stats:', error)
+    return res.status(500).json({
+      error: 'Failed to get session binding statistics',
+      message: error.message
+    })
+  }
+})
+
+module.exports = router

src/routes/admin/concurrency.js

@@ -0,0 +1,313 @@
+/**
+ * 并发管理 API 路由
+ * 提供并发状态查看和手动清理功能
+ */
+
+const express = require('express')
+const router = express.Router()
+const redis = require('../../models/redis')
+const logger = require('../../utils/logger')
+const { authenticateAdmin } = require('../../middleware/auth')
+const { calculateWaitTimeStats } = require('../../utils/statsHelper')
+
+/**
+ * GET /admin/concurrency
+ * 获取所有并发状态
+ */
+router.get('/concurrency', authenticateAdmin, async (req, res) => {
+  try {
+    const status = await redis.getAllConcurrencyStatus()
+
+    // 为每个 API Key 获取排队计数
+    const statusWithQueue = await Promise.all(
+      status.map(async (s) => {
+        const queueCount = await redis.getConcurrencyQueueCount(s.apiKeyId)
+        return {
+          ...s,
+          queueCount
+        }
+      })
+    )
+
+    // 计算汇总统计
+    const summary = {
+      totalKeys: statusWithQueue.length,
+      totalActiveRequests: statusWithQueue.reduce((sum, s) => sum + s.activeCount, 0),
+      totalExpiredRequests: statusWithQueue.reduce((sum, s) => sum + s.expiredCount, 0),
+      totalQueuedRequests: statusWithQueue.reduce((sum, s) => sum + s.queueCount, 0)
+    }
+
+    res.json({
+      success: true,
+      summary,
+      concurrencyStatus: statusWithQueue
+    })
+  } catch (error) {
+    logger.error('❌ Failed to get concurrency status:', error)
+    res.status(500).json({
+      success: false,
+      error: 'Failed to get concurrency status',
+      message: error.message
+    })
+  }
+})
+
+/**
+ * GET /admin/concurrency-queue/stats
+ * 获取排队统计信息
+ */
+router.get('/concurrency-queue/stats', authenticateAdmin, async (req, res) => {
+  try {
+    // 获取所有有统计数据的 API Key
+    const statsKeys = await redis.scanConcurrencyQueueStatsKeys()
+    const queueKeys = await redis.scanConcurrencyQueueKeys()
+
+    // 合并所有相关的 API Key
+    const allApiKeyIds = [...new Set([...statsKeys, ...queueKeys])]
+
+    // 获取各 API Key 的详细统计
+    const perKeyStats = await Promise.all(
+      allApiKeyIds.map(async (apiKeyId) => {
+        const [queueCount, stats, waitTimes] = await Promise.all([
+          redis.getConcurrencyQueueCount(apiKeyId),
+          redis.getConcurrencyQueueStats(apiKeyId),
+          redis.getQueueWaitTimes(apiKeyId)
+        ])
+
+        return {
+          apiKeyId,
+          currentQueueCount: queueCount,
+          stats,
+          waitTimeStats: calculateWaitTimeStats(waitTimes)
+        }
+      })
+    )
+
+    // 获取全局等待时间统计
+    const globalWaitTimes = await redis.getGlobalQueueWaitTimes()
+    const globalWaitTimeStats = calculateWaitTimeStats(globalWaitTimes)
+
+    // 计算全局汇总
+    const globalStats = {
+      totalEntered: perKeyStats.reduce((sum, s) => sum + s.stats.entered, 0),
+      totalSuccess: perKeyStats.reduce((sum, s) => sum + s.stats.success, 0),
+      totalTimeout: perKeyStats.reduce((sum, s) => sum + s.stats.timeout, 0),
+      totalCancelled: perKeyStats.reduce((sum, s) => sum + s.stats.cancelled, 0),
+      totalSocketChanged: perKeyStats.reduce((sum, s) => sum + (s.stats.socket_changed || 0), 0),
+      totalRejectedOverload: perKeyStats.reduce(
+        (sum, s) => sum + (s.stats.rejected_overload || 0),
+        0
+      ),
+      currentTotalQueued: perKeyStats.reduce((sum, s) => sum + s.currentQueueCount, 0),
+      // 队列资源利用率指标
+      peakQueueSize:
+        perKeyStats.length > 0 ? Math.max(...perKeyStats.map((s) => s.currentQueueCount)) : 0,
+      avgQueueSize:
+        perKeyStats.length > 0
+          ? Math.round(
+              perKeyStats.reduce((sum, s) => sum + s.currentQueueCount, 0) / perKeyStats.length
+            )
+          : 0,
+      activeApiKeys: perKeyStats.filter((s) => s.currentQueueCount > 0).length
+    }
+
+    // 计算成功率
+    if (globalStats.totalEntered > 0) {
+      globalStats.successRate = Math.round(
+        (globalStats.totalSuccess / globalStats.totalEntered) * 100
+      )
+      globalStats.timeoutRate = Math.round(
+        (globalStats.totalTimeout / globalStats.totalEntered) * 100
+      )
+      globalStats.cancelledRate = Math.round(
+        (globalStats.totalCancelled / globalStats.totalEntered) * 100
+      )
+    }
+
+    // 从全局等待时间统计中提取关键指标
+    if (globalWaitTimeStats) {
+      globalStats.avgWaitTimeMs = globalWaitTimeStats.avg
+      globalStats.p50WaitTimeMs = globalWaitTimeStats.p50
+      globalStats.p90WaitTimeMs = globalWaitTimeStats.p90
+      globalStats.p99WaitTimeMs = globalWaitTimeStats.p99
+      // 多实例采样策略标记（详见 design.md Decision 9）
+      // 全局 P90 仅用于可视化和监控，不用于系统决策
+      // 健康检查使用 API Key 级别的 P90（每 Key 独立采样）
+      globalWaitTimeStats.globalP90ForVisualizationOnly = true
+    }
+
+    res.json({
+      success: true,
+      globalStats,
+      globalWaitTimeStats,
+      perKeyStats
+    })
+  } catch (error) {
+    logger.error('❌ Failed to get queue stats:', error)
+    res.status(500).json({
+      success: false,
+      error: 'Failed to get queue stats',
+      message: error.message
+    })
+  }
+})
+
+/**
+ * DELETE /admin/concurrency-queue/:apiKeyId
+ * 清理特定 API Key 的排队计数
+ */
+router.delete('/concurrency-queue/:apiKeyId', authenticateAdmin, async (req, res) => {
+  try {
+    const { apiKeyId } = req.params
+    await redis.clearConcurrencyQueue(apiKeyId)
+
+    logger.warn(`🧹 Admin ${req.admin?.username || 'unknown'} cleared queue for key ${apiKeyId}`)
+
+    res.json({
+      success: true,
+      message: `Successfully cleared queue for API key ${apiKeyId}`
+    })
+  } catch (error) {
+    logger.error(`❌ Failed to clear queue for ${req.params.apiKeyId}:`, error)
+    res.status(500).json({
+      success: false,
+      error: 'Failed to clear queue',
+      message: error.message
+    })
+  }
+})
+
+/**
+ * DELETE /admin/concurrency-queue
+ * 清理所有排队计数
+ */
+router.delete('/concurrency-queue', authenticateAdmin, async (req, res) => {
+  try {
+    const cleared = await redis.clearAllConcurrencyQueues()
+
+    logger.warn(`🧹 Admin ${req.admin?.username || 'unknown'} cleared ALL queues`)
+
+    res.json({
+      success: true,
+      message: 'Successfully cleared all queues',
+      cleared
+    })
+  } catch (error) {
+    logger.error('❌ Failed to clear all queues:', error)
+    res.status(500).json({
+      success: false,
+      error: 'Failed to clear all queues',
+      message: error.message
+    })
+  }
+})
+
+/**
+ * GET /admin/concurrency/:apiKeyId
+ * 获取特定 API Key 的并发状态详情
+ */
+router.get('/concurrency/:apiKeyId', authenticateAdmin, async (req, res) => {
+  try {
+    const { apiKeyId } = req.params
+    const status = await redis.getConcurrencyStatus(apiKeyId)
+    const queueCount = await redis.getConcurrencyQueueCount(apiKeyId)
+
+    res.json({
+      success: true,
+      concurrencyStatus: {
+        ...status,
+        queueCount
+      }
+    })
+  } catch (error) {
+    logger.error(`❌ Failed to get concurrency status for ${req.params.apiKeyId}:`, error)
+    res.status(500).json({
+      success: false,
+      error: 'Failed to get concurrency status',
+      message: error.message
+    })
+  }
+})
+
+/**
+ * DELETE /admin/concurrency/:apiKeyId
+ * 强制清理特定 API Key 的并发计数
+ */
+router.delete('/concurrency/:apiKeyId', authenticateAdmin, async (req, res) => {
+  try {
+    const { apiKeyId } = req.params
+    const result = await redis.forceClearConcurrency(apiKeyId)
+
+    logger.warn(
+      `🧹 Admin ${req.admin?.username || 'unknown'} force cleared concurrency for key ${apiKeyId}`
+    )
+
+    res.json({
+      success: true,
+      message: `Successfully cleared concurrency for API key ${apiKeyId}`,
+      result
+    })
+  } catch (error) {
+    logger.error(`❌ Failed to clear concurrency for ${req.params.apiKeyId}:`, error)
+    res.status(500).json({
+      success: false,
+      error: 'Failed to clear concurrency',
+      message: error.message
+    })
+  }
+})
+
+/**
+ * DELETE /admin/concurrency
+ * 强制清理所有并发计数
+ */
+router.delete('/concurrency', authenticateAdmin, async (req, res) => {
+  try {
+    const result = await redis.forceClearAllConcurrency()
+
+    logger.warn(`🧹 Admin ${req.admin?.username || 'unknown'} force cleared ALL concurrency`)
+
+    res.json({
+      success: true,
+      message: 'Successfully cleared all concurrency',
+      result
+    })
+  } catch (error) {
+    logger.error('❌ Failed to clear all concurrency:', error)
+    res.status(500).json({
+      success: false,
+      error: 'Failed to clear all concurrency',
+      message: error.message
+    })
+  }
+})
+
+/**
+ * POST /admin/concurrency/cleanup
+ * 清理过期的并发条目（不影响活跃请求）
+ */
+router.post('/concurrency/cleanup', authenticateAdmin, async (req, res) => {
+  try {
+    const { apiKeyId } = req.body
+    const result = await redis.cleanupExpiredConcurrency(apiKeyId || null)
+
+    logger.info(`🧹 Admin ${req.admin?.username || 'unknown'} cleaned up expired concurrency`)
+
+    res.json({
+      success: true,
+      message: apiKeyId
+        ? `Successfully cleaned up expired concurrency for API key ${apiKeyId}`
+        : 'Successfully cleaned up all expired concurrency',
+      result
+    })
+  } catch (error) {
+    logger.error('❌ Failed to cleanup expired concurrency:', error)
+    res.status(500).json({
+      success: false,
+      error: 'Failed to cleanup expired concurrency',
+      message: error.message
+    })
+  }
+})
+
+module.exports = router

src/routes/admin/dashboard.js

@@ -6,13 +6,11 @@ const bedrockAccountService = require('../../services/bedrockAccountService')
 const ccrAccountService = require('../../services/ccrAccountService')
 const geminiAccountService = require('../../services/geminiAccountService')
 const droidAccountService = require('../../services/droidAccountService')
-const openaiAccountService = require('../../services/openaiAccountService')
 const openaiResponsesAccountService = require('../../services/openaiResponsesAccountService')
 const redis = require('../../models/redis')
 const { authenticateAdmin } = require('../../middleware/auth')
 const logger = require('../../utils/logger')
 const CostCalculator = require('../../utils/costCalculator')
-const pricingService = require('../../services/pricingService')
 const config = require('../../../config/config')
 
 const router = express.Router()

src/routes/admin/geminiAccounts.js

@@ -11,14 +11,19 @@ const { formatAccountExpiry, mapExpiryField } = require('./utils')
 const router = express.Router()
 
 // 🤖 Gemini OAuth 账户管理
+function getDefaultRedirectUri(oauthProvider) {
+  if (oauthProvider === 'antigravity') {
+    return process.env.ANTIGRAVITY_OAUTH_REDIRECT_URI || 'http://localhost:45462'
+  }
+  return process.env.GEMINI_OAUTH_REDIRECT_URI || 'https://codeassist.google.com/authcode'
+}
 
 // 生成 Gemini OAuth 授权 URL
 router.post('/generate-auth-url', authenticateAdmin, async (req, res) => {
   try {
-    const { state, proxy } = req.body // 接收代理配置
+    const { state, proxy, oauthProvider } = req.body // 接收代理配置与OAuth Provider
 
-    // 使用新的 codeassist.google.com 回调地址
-    const redirectUri = 'https://codeassist.google.com/authcode'
+    const redirectUri = getDefaultRedirectUri(oauthProvider)
 
     logger.info(`Generating Gemini OAuth URL with redirect_uri: ${redirectUri}`)
 
@@ -26,8 +31,9 @@ router.post('/generate-auth-url', authenticateAdmin, async (req, res) => {
       authUrl,
       state: authState,
       codeVerifier,
-      redirectUri: finalRedirectUri
-    } = await geminiAccountService.generateAuthUrl(state, redirectUri, proxy)
+      redirectUri: finalRedirectUri,
+      oauthProvider: resolvedOauthProvider
+    } = await geminiAccountService.generateAuthUrl(state, redirectUri, proxy, oauthProvider)
 
     // 创建 OAuth 会话，包含 codeVerifier 和代理配置
     const sessionId = authState
@@ -37,6 +43,7 @@ router.post('/generate-auth-url', authenticateAdmin, async (req, res) => {
       redirectUri: finalRedirectUri,
       codeVerifier, // 保存 PKCE code verifier
       proxy: proxy || null, // 保存代理配置
+      oauthProvider: resolvedOauthProvider,
       createdAt: new Date().toISOString()
     })
 
@@ -45,7 +52,8 @@ router.post('/generate-auth-url', authenticateAdmin, async (req, res) => {
       success: true,
       data: {
         authUrl,
-        sessionId
+        sessionId,
+        oauthProvider: resolvedOauthProvider
       }
     })
   } catch (error) {
@@ -80,13 +88,14 @@ router.post('/poll-auth-status', authenticateAdmin, async (req, res) => {
 // 交换 Gemini 授权码
 router.post('/exchange-code', authenticateAdmin, async (req, res) => {
   try {
-    const { code, sessionId, proxy: requestProxy } = req.body
+    const { code, sessionId, proxy: requestProxy, oauthProvider } = req.body
+    let resolvedOauthProvider = oauthProvider
 
     if (!code) {
       return res.status(400).json({ error: 'Authorization code is required' })
     }
 
-    let redirectUri = 'https://codeassist.google.com/authcode'
+    let redirectUri = getDefaultRedirectUri(resolvedOauthProvider)
     let codeVerifier = null
     let proxyConfig = null
 
@@ -97,11 +106,16 @@ router.post('/exchange-code', authenticateAdmin, async (req, res) => {
         const {
           redirectUri: sessionRedirectUri,
           codeVerifier: sessionCodeVerifier,
-          proxy
+          proxy,
+          oauthProvider: sessionOauthProvider
         } = sessionData
         redirectUri = sessionRedirectUri || redirectUri
         codeVerifier = sessionCodeVerifier
         proxyConfig = proxy // 获取代理配置
+        if (!resolvedOauthProvider && sessionOauthProvider) {
+          // 会话里保存的 provider 仅作为兜底
+          resolvedOauthProvider = sessionOauthProvider
+        }
         logger.info(
           `Using session redirect_uri: ${redirectUri}, has codeVerifier: ${!!codeVerifier}, has proxy from session: ${!!proxyConfig}`
         )
@@ -120,7 +134,8 @@ router.post('/exchange-code', authenticateAdmin, async (req, res) => {
       code,
       redirectUri,
       codeVerifier,
-      proxyConfig // 传递代理配置
+      proxyConfig, // 传递代理配置
+      resolvedOauthProvider
     )
 
     // 清理 OAuth 会话
@@ -129,7 +144,7 @@ router.post('/exchange-code', authenticateAdmin, async (req, res) => {
     }
 
     logger.success('✅ Successfully exchanged Gemini authorization code')
-    return res.json({ success: true, data: { tokens } })
+    return res.json({ success: true, data: { tokens, oauthProvider: resolvedOauthProvider } })
   } catch (error) {
     logger.error('❌ Failed to exchange Gemini authorization code:', error)
     return res.status(500).json({ error: 'Failed to exchange code', message: error.message })

src/routes/admin/index.js

@@ -21,7 +21,11 @@ const openaiResponsesAccountsRoutes = require('./openaiResponsesAccounts')
 const droidAccountsRoutes = require('./droidAccounts')
 const dashboardRoutes = require('./dashboard')
 const usageStatsRoutes = require('./usageStats')
+const accountBalanceRoutes = require('./accountBalance')
 const systemRoutes = require('./system')
+const concurrencyRoutes = require('./concurrency')
+const claudeRelayConfigRoutes = require('./claudeRelayConfig')
+const syncRoutes = require('./sync')
 
 // 挂载所有子路由
 // 使用完整路径的模块（直接挂载到根路径）
@@ -34,7 +38,11 @@ router.use('/', openaiResponsesAccountsRoutes)
 router.use('/', droidAccountsRoutes)
 router.use('/', dashboardRoutes)
 router.use('/', usageStatsRoutes)
+router.use('/', accountBalanceRoutes)
 router.use('/', systemRoutes)
+router.use('/', concurrencyRoutes)
+router.use('/', claudeRelayConfigRoutes)
+router.use('/', syncRoutes)
 
 // 使用相对路径的模块（需要指定基础路径前缀）
 router.use('/account-groups', accountGroupsRoutes)

src/routes/admin/sync.js

@@ -0,0 +1,460 @@
+/**
+ * Admin Routes - Sync / Export (for migration)
+ * Exports account data (including secrets) for safe server-to-server syncing.
+ */
+
+const express = require('express')
+const router = express.Router()
+
+const { authenticateAdmin } = require('../../middleware/auth')
+const redis = require('../../models/redis')
+const claudeAccountService = require('../../services/claudeAccountService')
+const claudeConsoleAccountService = require('../../services/claudeConsoleAccountService')
+const openaiAccountService = require('../../services/openaiAccountService')
+const openaiResponsesAccountService = require('../../services/openaiResponsesAccountService')
+const logger = require('../../utils/logger')
+
+function toBool(value, defaultValue = false) {
+  if (value === undefined || value === null || value === '') {
+    return defaultValue
+  }
+  if (value === true || value === 'true') {
+    return true
+  }
+  if (value === false || value === 'false') {
+    return false
+  }
+  return defaultValue
+}
+
+function normalizeProxy(proxy) {
+  if (!proxy || typeof proxy !== 'object') {
+    return null
+  }
+
+  const protocol = proxy.protocol || proxy.type || proxy.scheme || ''
+  const host = proxy.host || ''
+  const port = Number(proxy.port || 0)
+
+  if (!protocol || !host || !Number.isFinite(port) || port <= 0) {
+    return null
+  }
+
+  return {
+    protocol: String(protocol),
+    host: String(host),
+    port,
+    username: proxy.username ? String(proxy.username) : '',
+    password: proxy.password ? String(proxy.password) : ''
+  }
+}
+
+function buildModelMappingFromSupportedModels(supportedModels) {
+  if (!supportedModels) {
+    return null
+  }
+
+  if (Array.isArray(supportedModels)) {
+    const mapping = {}
+    for (const model of supportedModels) {
+      if (typeof model === 'string' && model.trim()) {
+        mapping[model.trim()] = model.trim()
+      }
+    }
+    return Object.keys(mapping).length ? mapping : null
+  }
+
+  if (typeof supportedModels === 'object') {
+    const mapping = {}
+    for (const [from, to] of Object.entries(supportedModels)) {
+      if (typeof from === 'string' && typeof to === 'string' && from.trim() && to.trim()) {
+        mapping[from.trim()] = to.trim()
+      }
+    }
+    return Object.keys(mapping).length ? mapping : null
+  }
+
+  return null
+}
+
+function safeParseJson(raw, fallback = null) {
+  if (!raw || typeof raw !== 'string') {
+    return fallback
+  }
+  try {
+    return JSON.parse(raw)
+  } catch (_) {
+    return fallback
+  }
+}
+
+// Export accounts for migration (includes secrets).
+// GET /admin/sync/export-accounts?include_secrets=true
+router.get('/sync/export-accounts', authenticateAdmin, async (req, res) => {
+  try {
+    const includeSecrets = toBool(req.query.include_secrets, false)
+    if (!includeSecrets) {
+      return res.status(400).json({
+        success: false,
+        error: 'include_secrets_required',
+        message: 'Set include_secrets=true to export secrets'
+      })
+    }
+
+    // ===== Claude official OAuth / Setup Token accounts =====
+    const rawClaudeAccounts = await redis.getAllClaudeAccounts()
+    const claudeAccounts = rawClaudeAccounts.map((account) => {
+      // Backward compatible extraction: prefer individual fields, fallback to claudeAiOauth JSON blob.
+      let decryptedClaudeAiOauth = null
+      if (account.claudeAiOauth) {
+        try {
+          const raw = claudeAccountService._decryptSensitiveData(account.claudeAiOauth)
+          decryptedClaudeAiOauth = raw ? JSON.parse(raw) : null
+        } catch (_) {
+          decryptedClaudeAiOauth = null
+        }
+      }
+
+      const rawScopes =
+        account.scopes && account.scopes.trim()
+          ? account.scopes
+          : decryptedClaudeAiOauth?.scopes
+            ? decryptedClaudeAiOauth.scopes.join(' ')
+            : ''
+
+      const scopes = rawScopes && rawScopes.trim() ? rawScopes.trim().split(' ') : []
+      const isOAuth = scopes.includes('user:profile') && scopes.includes('user:inference')
+      const authType = isOAuth ? 'oauth' : 'setup-token'
+
+      const accessToken =
+        account.accessToken && String(account.accessToken).trim()
+          ? claudeAccountService._decryptSensitiveData(account.accessToken)
+          : decryptedClaudeAiOauth?.accessToken || ''
+
+      const refreshToken =
+        account.refreshToken && String(account.refreshToken).trim()
+          ? claudeAccountService._decryptSensitiveData(account.refreshToken)
+          : decryptedClaudeAiOauth?.refreshToken || ''
+
+      let expiresAt = null
+      const expiresAtMs = Number.parseInt(account.expiresAt, 10)
+      if (Number.isFinite(expiresAtMs) && expiresAtMs > 0) {
+        expiresAt = new Date(expiresAtMs).toISOString()
+      } else if (decryptedClaudeAiOauth?.expiresAt) {
+        try {
+          expiresAt = new Date(Number(decryptedClaudeAiOauth.expiresAt)).toISOString()
+        } catch (_) {
+          expiresAt = null
+        }
+      }
+
+      const proxy = account.proxy ? normalizeProxy(safeParseJson(account.proxy)) : null
+
+      // 🔧 Parse subscriptionInfo to extract org_uuid and account_uuid
+      let orgUuid = null
+      let accountUuid = null
+      if (account.subscriptionInfo) {
+        try {
+          const subscriptionInfo = JSON.parse(account.subscriptionInfo)
+          orgUuid = subscriptionInfo.organizationUuid || null
+          accountUuid = subscriptionInfo.accountUuid || null
+        } catch (_) {
+          // Ignore parse errors
+        }
+      }
+
+      // 🔧 Calculate expires_in from expires_at
+      let expiresIn = null
+      if (expiresAt) {
+        try {
+          const expiresAtTime = new Date(expiresAt).getTime()
+          const nowTime = Date.now()
+          const diffSeconds = Math.floor((expiresAtTime - nowTime) / 1000)
+          if (diffSeconds > 0) {
+            expiresIn = diffSeconds
+          }
+        } catch (_) {
+          // Ignore calculation errors
+        }
+      }
+      // 🔧 Use default expires_in if calculation failed (Anthropic OAuth: 8 hours)
+      if (!expiresIn && isOAuth) {
+        expiresIn = 28800 // 8 hours
+      }
+
+      const credentials = {
+        access_token: accessToken,
+        refresh_token: refreshToken || undefined,
+        expires_at: expiresAt || undefined,
+        expires_in: expiresIn || undefined,
+        scope: scopes.join(' ') || undefined,
+        token_type: 'Bearer'
+      }
+      // 🔧 Add auth info as top-level credentials fields
+      if (orgUuid) {
+        credentials.org_uuid = orgUuid
+      }
+      if (accountUuid) {
+        credentials.account_uuid = accountUuid
+      }
+
+      // 🔧 Store complete original CRS data in extra
+      const extra = {
+        crs_account_id: account.id,
+        crs_kind: 'claude-account',
+        crs_id: account.id,
+        crs_name: account.name,
+        crs_description: account.description || '',
+        crs_platform: account.platform || 'claude',
+        crs_auth_type: authType,
+        crs_is_active: account.isActive === 'true',
+        crs_schedulable: account.schedulable !== 'false',
+        crs_priority: Number.parseInt(account.priority, 10) || 50,
+        crs_status: account.status || 'active',
+        crs_scopes: scopes,
+        crs_subscription_info: account.subscriptionInfo || undefined
+      }
+
+      return {
+        kind: 'claude-account',
+        id: account.id,
+        name: account.name,
+        description: account.description || '',
+        platform: account.platform || 'claude',
+        authType,
+        isActive: account.isActive === 'true',
+        schedulable: account.schedulable !== 'false',
+        priority: Number.parseInt(account.priority, 10) || 50,
+        status: account.status || 'active',
+        proxy,
+        credentials,
+        extra
+      }
+    })
+
+    // ===== Claude Console API Key accounts =====
+    const claudeConsoleSummaries = await claudeConsoleAccountService.getAllAccounts()
+    const claudeConsoleAccounts = []
+    for (const summary of claudeConsoleSummaries) {
+      const full = await claudeConsoleAccountService.getAccount(summary.id)
+      if (!full) {
+        continue
+      }
+
+      const proxy = normalizeProxy(full.proxy)
+      const modelMapping = buildModelMappingFromSupportedModels(full.supportedModels)
+
+      const credentials = {
+        api_key: full.apiKey,
+        base_url: full.apiUrl
+      }
+
+      if (modelMapping) {
+        credentials.model_mapping = modelMapping
+      }
+
+      if (full.userAgent) {
+        credentials.user_agent = full.userAgent
+      }
+
+      claudeConsoleAccounts.push({
+        kind: 'claude-console-account',
+        id: full.id,
+        name: full.name,
+        description: full.description || '',
+        platform: full.platform || 'claude-console',
+        isActive: full.isActive === true,
+        schedulable: full.schedulable !== false,
+        priority: Number.parseInt(full.priority, 10) || 50,
+        status: full.status || 'active',
+        proxy,
+        maxConcurrentTasks: Number.parseInt(full.maxConcurrentTasks, 10) || 0,
+        credentials,
+        extra: {
+          crs_account_id: full.id,
+          crs_kind: 'claude-console-account',
+          crs_id: full.id,
+          crs_name: full.name,
+          crs_description: full.description || '',
+          crs_platform: full.platform || 'claude-console',
+          crs_is_active: full.isActive === true,
+          crs_schedulable: full.schedulable !== false,
+          crs_priority: Number.parseInt(full.priority, 10) || 50,
+          crs_status: full.status || 'active'
+        }
+      })
+    }
+
+    // ===== OpenAI OAuth accounts =====
+    const openaiOAuthAccounts = []
+    {
+      const client = redis.getClientSafe()
+      const openaiKeys = await client.keys('openai:account:*')
+      for (const key of openaiKeys) {
+        const id = key.split(':').slice(2).join(':')
+        const account = await openaiAccountService.getAccount(id)
+        if (!account) {
+          continue
+        }
+
+        const accessToken = account.accessToken
+          ? openaiAccountService.decrypt(account.accessToken)
+          : ''
+        if (!accessToken) {
+          // Skip broken/legacy records without decryptable token
+          continue
+        }
+
+        const scopes =
+          account.scopes && typeof account.scopes === 'string' && account.scopes.trim()
+            ? account.scopes.trim().split(' ')
+            : []
+
+        const proxy = normalizeProxy(account.proxy)
+
+        // 🔧 Calculate expires_in from expires_at
+        let expiresIn = null
+        if (account.expiresAt) {
+          try {
+            const expiresAtTime = new Date(account.expiresAt).getTime()
+            const nowTime = Date.now()
+            const diffSeconds = Math.floor((expiresAtTime - nowTime) / 1000)
+            if (diffSeconds > 0) {
+              expiresIn = diffSeconds
+            }
+          } catch (_) {
+            // Ignore calculation errors
+          }
+        }
+        // 🔧 Use default expires_in if calculation failed (OpenAI OAuth: 10 days)
+        if (!expiresIn) {
+          expiresIn = 864000 // 10 days
+        }
+
+        const credentials = {
+          access_token: accessToken,
+          refresh_token: account.refreshToken || undefined,
+          id_token: account.idToken || undefined,
+          expires_at: account.expiresAt || undefined,
+          expires_in: expiresIn || undefined,
+          scope: scopes.join(' ') || undefined,
+          token_type: 'Bearer'
+        }
+        // 🔧 Add auth info as top-level credentials fields
+        if (account.accountId) {
+          credentials.chatgpt_account_id = account.accountId
+        }
+        if (account.chatgptUserId) {
+          credentials.chatgpt_user_id = account.chatgptUserId
+        }
+        if (account.organizationId) {
+          credentials.organization_id = account.organizationId
+        }
+
+        // 🔧 Store complete original CRS data in extra
+        const extra = {
+          crs_account_id: account.id,
+          crs_kind: 'openai-oauth-account',
+          crs_id: account.id,
+          crs_name: account.name,
+          crs_description: account.description || '',
+          crs_platform: account.platform || 'openai',
+          crs_is_active: account.isActive === 'true',
+          crs_schedulable: account.schedulable !== 'false',
+          crs_priority: Number.parseInt(account.priority, 10) || 50,
+          crs_status: account.status || 'active',
+          crs_scopes: scopes,
+          crs_email: account.email || undefined,
+          crs_chatgpt_account_id: account.accountId || undefined,
+          crs_chatgpt_user_id: account.chatgptUserId || undefined,
+          crs_organization_id: account.organizationId || undefined
+        }
+
+        openaiOAuthAccounts.push({
+          kind: 'openai-oauth-account',
+          id: account.id,
+          name: account.name,
+          description: account.description || '',
+          platform: account.platform || 'openai',
+          authType: 'oauth',
+          isActive: account.isActive === 'true',
+          schedulable: account.schedulable !== 'false',
+          priority: Number.parseInt(account.priority, 10) || 50,
+          status: account.status || 'active',
+          proxy,
+          credentials,
+          extra
+        })
+      }
+    }
+
+    // ===== OpenAI Responses API Key accounts =====
+    const openaiResponsesAccounts = []
+    const client = redis.getClientSafe()
+    const openaiResponseKeys = await client.keys('openai_responses_account:*')
+    for (const key of openaiResponseKeys) {
+      const id = key.split(':').slice(1).join(':')
+      const full = await openaiResponsesAccountService.getAccount(id)
+      if (!full) {
+        continue
+      }
+
+      const proxy = normalizeProxy(full.proxy)
+
+      const credentials = {
+        api_key: full.apiKey,
+        base_url: full.baseApi
+      }
+
+      if (full.userAgent) {
+        credentials.user_agent = full.userAgent
+      }
+
+      openaiResponsesAccounts.push({
+        kind: 'openai-responses-account',
+        id: full.id,
+        name: full.name,
+        description: full.description || '',
+        platform: full.platform || 'openai-responses',
+        isActive: full.isActive === 'true',
+        schedulable: full.schedulable !== 'false',
+        priority: Number.parseInt(full.priority, 10) || 50,
+        status: full.status || 'active',
+        proxy,
+        credentials,
+        extra: {
+          crs_account_id: full.id,
+          crs_kind: 'openai-responses-account',
+          crs_id: full.id,
+          crs_name: full.name,
+          crs_description: full.description || '',
+          crs_platform: full.platform || 'openai-responses',
+          crs_is_active: full.isActive === 'true',
+          crs_schedulable: full.schedulable !== 'false',
+          crs_priority: Number.parseInt(full.priority, 10) || 50,
+          crs_status: full.status || 'active'
+        }
+      })
+    }
+
+    return res.json({
+      success: true,
+      data: {
+        exportedAt: new Date().toISOString(),
+        claudeAccounts,
+        claudeConsoleAccounts,
+        openaiOAuthAccounts,
+        openaiResponsesAccounts
+      }
+    })
+  } catch (error) {
+    logger.error('❌ Failed to export accounts for sync:', error)
+    return res.status(500).json({
+      success: false,
+      error: 'export_failed',
+      message: error.message
+    })
+  }
+})
+
+module.exports = router

src/routes/admin/usageStats.js

@@ -1,11 +1,14 @@
 const express = require('express')
 const apiKeyService = require('../../services/apiKeyService')
+const ccrAccountService = require('../../services/ccrAccountService')
 const claudeAccountService = require('../../services/claudeAccountService')
 const claudeConsoleAccountService = require('../../services/claudeConsoleAccountService')
 const geminiAccountService = require('../../services/geminiAccountService')
+const geminiApiAccountService = require('../../services/geminiApiAccountService')
 const openaiAccountService = require('../../services/openaiAccountService')
 const openaiResponsesAccountService = require('../../services/openaiResponsesAccountService')
 const droidAccountService = require('../../services/droidAccountService')
+const bedrockAccountService = require('../../services/bedrockAccountService')
 const redis = require('../../models/redis')
 const { authenticateAdmin } = require('../../middleware/auth')
 const logger = require('../../utils/logger')
@@ -14,6 +17,67 @@ const pricingService = require('../../services/pricingService')
 
 const router = express.Router()
 
+const accountTypeNames = {
+  claude: 'Claude官方',
+  'claude-console': 'Claude Console',
+  ccr: 'Claude Console Relay',
+  openai: 'OpenAI',
+  'openai-responses': 'OpenAI Responses',
+  gemini: 'Gemini',
+  'gemini-api': 'Gemini API',
+  droid: 'Droid',
+  bedrock: 'AWS Bedrock',
+  unknown: '未知渠道'
+}
+
+const resolveAccountByPlatform = async (accountId, platform) => {
+  const serviceMap = {
+    claude: claudeAccountService,
+    'claude-console': claudeConsoleAccountService,
+    gemini: geminiAccountService,
+    'gemini-api': geminiApiAccountService,
+    openai: openaiAccountService,
+    'openai-responses': openaiResponsesAccountService,
+    droid: droidAccountService,
+    ccr: ccrAccountService,
+    bedrock: bedrockAccountService
+  }
+
+  if (platform && serviceMap[platform]) {
+    try {
+      const account = await serviceMap[platform].getAccount(accountId)
+      if (account) {
+        return { ...account, platform }
+      }
+    } catch (error) {
+      logger.debug(`⚠️ Failed to get account ${accountId} from ${platform}: ${error.message}`)
+    }
+  }
+
+  for (const [platformName, service] of Object.entries(serviceMap)) {
+    try {
+      const account = await service.getAccount(accountId)
+      if (account) {
+        return { ...account, platform: platformName }
+      }
+    } catch (error) {
+      logger.debug(`⚠️ Failed to get account ${accountId} from ${platformName}: ${error.message}`)
+    }
+  }
+
+  return null
+}
+
+const getApiKeyName = async (keyId) => {
+  try {
+    const keyData = await redis.getApiKey(keyId)
+    return keyData?.name || keyData?.label || keyId
+  } catch (error) {
+    logger.debug(`⚠️ Failed to get API key name for ${keyId}: ${error.message}`)
+    return keyId
+  }
+}
+
 // 📊 账户使用统计
 
 // 获取所有账户的使用统计
@@ -100,7 +164,8 @@ router.get('/accounts/:accountId/usage-history', authenticateAdmin, async (req,
       'openai-responses',
       'gemini',
       'gemini-api',
-      'droid'
+      'droid',
+      'bedrock'
     ]
     if (!allowedPlatforms.includes(platform)) {
       return res.status(400).json({
@@ -113,7 +178,8 @@ router.get('/accounts/:accountId/usage-history', authenticateAdmin, async (req,
       openai: 'openai',
       'openai-responses': 'openai-responses',
       'gemini-api': 'gemini-api',
-      droid: 'droid'
+      droid: 'droid',
+      bedrock: 'bedrock'
     }
 
     const fallbackModelMap = {
@@ -123,7 +189,8 @@ router.get('/accounts/:accountId/usage-history', authenticateAdmin, async (req,
       'openai-responses': 'gpt-4o-mini-2024-07-18',
       gemini: 'gemini-1.5-flash',
       'gemini-api': 'gemini-2.0-flash',
-      droid: 'unknown'
+      droid: 'unknown',
+      bedrock: 'us.anthropic.claude-3-5-sonnet-20241022-v2:0'
     }
 
     // 获取账户信息以获取创建时间
@@ -148,13 +215,17 @@ router.get('/accounts/:accountId/usage-history', authenticateAdmin, async (req,
           accountData = await geminiAccountService.getAccount(accountId)
           break
         case 'gemini-api': {
-          const geminiApiAccountService = require('../../services/geminiApiAccountService')
           accountData = await geminiApiAccountService.getAccount(accountId)
           break
         }
         case 'droid':
           accountData = await droidAccountService.getAccount(accountId)
           break
+        case 'bedrock': {
+          const result = await bedrockAccountService.getAccount(accountId)
+          accountData = result?.success ? result.data : null
+          break
+        }
       }
 
       if (accountData && accountData.createdAt) {
@@ -369,7 +440,9 @@ router.get('/usage-trend', authenticateAdmin, async (req, res) => {
         logger.info(`  endDate (raw): ${endDate}`)
         logger.info(`  startTime (parsed): ${startTime.toISOString()}`)
         logger.info(`  endTime (parsed): ${endTime.toISOString()}`)
-        logger.info(`  System timezone offset: ${require('../../../config/config').system.timezoneOffset || 8}`)
+        logger.info(
+          `  System timezone offset: ${require('../../../config/config').system.timezoneOffset || 8}`
+        )
       } else {
         // 默认最近24小时
         endTime = new Date()
@@ -820,7 +893,7 @@ router.get('/account-usage-trend', authenticateAdmin, async (req, res) => {
   try {
     const { granularity = 'day', group = 'claude', days = 7, startDate, endDate } = req.query
 
-    const allowedGroups = ['claude', 'openai', 'gemini', 'droid']
+    const allowedGroups = ['claude', 'openai', 'gemini', 'droid', 'bedrock']
     if (!allowedGroups.includes(group)) {
       return res.status(400).json({
         success: false,
@@ -832,7 +905,8 @@ router.get('/account-usage-trend', authenticateAdmin, async (req, res) => {
       claude: 'Claude账户',
       openai: 'OpenAI账户',
       gemini: 'Gemini账户',
-      droid: 'Droid账户'
+      droid: 'Droid账户',
+      bedrock: 'Bedrock账户'
     }
 
     // 拉取各平台账号列表
@@ -890,7 +964,6 @@ router.get('/account-usage-trend', authenticateAdmin, async (req, res) => {
         })
       ]
     } else if (group === 'gemini') {
-      const geminiApiAccountService = require('../../services/geminiApiAccountService')
       const [geminiAccounts, geminiApiAccounts] = await Promise.all([
         geminiAccountService.getAllAccounts(),
         geminiApiAccountService.getAllAccounts(true)
@@ -927,6 +1000,18 @@ router.get('/account-usage-trend', authenticateAdmin, async (req, res) => {
           platform: 'droid'
         }
       })
+    } else if (group === 'bedrock') {
+      const result = await bedrockAccountService.getAllAccounts()
+      const bedrockAccounts = result?.success ? result.data : []
+      accounts = bedrockAccounts.map((account) => {
+        const id = String(account.id || '')
+        const shortId = id ? id.slice(0, 8) : '未知'
+        return {
+          id,
+          name: account.name || `Bedrock账号 ${shortId}`,
+          platform: 'bedrock'
+        }
+      })
     }
 
     if (!accounts || accounts.length === 0) {
@@ -1818,4 +1903,628 @@ router.get('/usage-costs', authenticateAdmin, async (req, res) => {
   }
 })
 
+// 获取 API Key 的请求记录时间线
+router.get('/api-keys/:keyId/usage-records', authenticateAdmin, async (req, res) => {
+  try {
+    const { keyId } = req.params
+    const {
+      page = 1,
+      pageSize = 50,
+      startDate,
+      endDate,
+      model,
+      accountId,
+      sortOrder = 'desc'
+    } = req.query
+
+    const pageNumber = Math.max(parseInt(page, 10) || 1, 1)
+    const pageSizeNumber = Math.min(Math.max(parseInt(pageSize, 10) || 50, 1), 200)
+    const normalizedSortOrder = sortOrder === 'asc' ? 'asc' : 'desc'
+
+    const startTime = startDate ? new Date(startDate) : null
+    const endTime = endDate ? new Date(endDate) : null
+
+    if (
+      (startDate && Number.isNaN(startTime?.getTime())) ||
+      (endDate && Number.isNaN(endTime?.getTime()))
+    ) {
+      return res.status(400).json({ success: false, error: 'Invalid date range' })
+    }
+
+    if (startTime && endTime && startTime > endTime) {
+      return res
+        .status(400)
+        .json({ success: false, error: 'Start date must be before or equal to end date' })
+    }
+
+    const apiKeyInfo = await redis.getApiKey(keyId)
+    if (!apiKeyInfo || Object.keys(apiKeyInfo).length === 0) {
+      return res.status(404).json({ success: false, error: 'API key not found' })
+    }
+
+    const rawRecords = await redis.getUsageRecords(keyId, 5000)
+
+    const accountServices = [
+      { type: 'claude', getter: (id) => claudeAccountService.getAccount(id) },
+      { type: 'claude-console', getter: (id) => claudeConsoleAccountService.getAccount(id) },
+      { type: 'ccr', getter: (id) => ccrAccountService.getAccount(id) },
+      { type: 'openai', getter: (id) => openaiAccountService.getAccount(id) },
+      { type: 'openai-responses', getter: (id) => openaiResponsesAccountService.getAccount(id) },
+      { type: 'gemini', getter: (id) => geminiAccountService.getAccount(id) },
+      { type: 'gemini-api', getter: (id) => geminiApiAccountService.getAccount(id) },
+      { type: 'droid', getter: (id) => droidAccountService.getAccount(id) }
+    ]
+
+    const accountCache = new Map()
+    const resolveAccountInfo = async (id, type) => {
+      if (!id) {
+        return null
+      }
+
+      const cacheKey = `${type || 'any'}:${id}`
+      if (accountCache.has(cacheKey)) {
+        return accountCache.get(cacheKey)
+      }
+
+      let servicesToTry = type
+        ? accountServices.filter((svc) => svc.type === type)
+        : accountServices
+
+      // 若渠道改名或传入未知类型，回退尝试全量服务，避免漏解析历史账号
+      if (!servicesToTry.length) {
+        servicesToTry = accountServices
+      }
+
+      for (const service of servicesToTry) {
+        try {
+          const account = await service.getter(id)
+          if (account) {
+            const info = {
+              id,
+              name: account.name || account.email || id,
+              type: service.type,
+              status: account.status || account.isActive
+            }
+            accountCache.set(cacheKey, info)
+            return info
+          }
+        } catch (error) {
+          logger.debug(`⚠️ Failed to resolve account ${id} via ${service.type}: ${error.message}`)
+        }
+      }
+
+      accountCache.set(cacheKey, null)
+      return null
+    }
+
+    const toUsageObject = (record) => ({
+      input_tokens: record.inputTokens || 0,
+      output_tokens: record.outputTokens || 0,
+      cache_creation_input_tokens: record.cacheCreateTokens || 0,
+      cache_read_input_tokens: record.cacheReadTokens || 0,
+      cache_creation: record.cacheCreation || record.cache_creation || null
+    })
+
+    const withinRange = (record) => {
+      if (!record.timestamp) {
+        return false
+      }
+      const ts = new Date(record.timestamp)
+      if (Number.isNaN(ts.getTime())) {
+        return false
+      }
+      if (startTime && ts < startTime) {
+        return false
+      }
+      if (endTime && ts > endTime) {
+        return false
+      }
+      return true
+    }
+
+    const filteredRecords = rawRecords.filter((record) => {
+      if (!withinRange(record)) {
+        return false
+      }
+      if (model && record.model !== model) {
+        return false
+      }
+      if (accountId && record.accountId !== accountId) {
+        return false
+      }
+      return true
+    })
+
+    filteredRecords.sort((a, b) => {
+      const aTime = new Date(a.timestamp).getTime()
+      const bTime = new Date(b.timestamp).getTime()
+      if (Number.isNaN(aTime) || Number.isNaN(bTime)) {
+        return 0
+      }
+      return normalizedSortOrder === 'asc' ? aTime - bTime : bTime - aTime
+    })
+
+    const summary = {
+      totalRequests: 0,
+      inputTokens: 0,
+      outputTokens: 0,
+      cacheCreateTokens: 0,
+      cacheReadTokens: 0,
+      totalTokens: 0,
+      totalCost: 0
+    }
+
+    const modelSet = new Set()
+    const accountOptionMap = new Map()
+    let earliestTimestamp = null
+    let latestTimestamp = null
+
+    for (const record of filteredRecords) {
+      const usage = toUsageObject(record)
+      const costData = CostCalculator.calculateCost(usage, record.model || 'unknown')
+      const computedCost =
+        typeof record.cost === 'number' ? record.cost : costData?.costs?.total || 0
+      const totalTokens =
+        record.totalTokens ||
+        usage.input_tokens +
+          usage.output_tokens +
+          usage.cache_creation_input_tokens +
+          usage.cache_read_input_tokens
+
+      summary.totalRequests += 1
+      summary.inputTokens += usage.input_tokens
+      summary.outputTokens += usage.output_tokens
+      summary.cacheCreateTokens += usage.cache_creation_input_tokens
+      summary.cacheReadTokens += usage.cache_read_input_tokens
+      summary.totalTokens += totalTokens
+      summary.totalCost += computedCost
+
+      if (record.model) {
+        modelSet.add(record.model)
+      }
+
+      if (record.accountId) {
+        const normalizedType = record.accountType || 'unknown'
+        if (!accountOptionMap.has(record.accountId)) {
+          accountOptionMap.set(record.accountId, {
+            id: record.accountId,
+            accountTypes: new Set([normalizedType])
+          })
+        } else {
+          accountOptionMap.get(record.accountId).accountTypes.add(normalizedType)
+        }
+      }
+
+      if (record.timestamp) {
+        const ts = new Date(record.timestamp)
+        if (!Number.isNaN(ts.getTime())) {
+          if (!earliestTimestamp || ts < earliestTimestamp) {
+            earliestTimestamp = ts
+          }
+          if (!latestTimestamp || ts > latestTimestamp) {
+            latestTimestamp = ts
+          }
+        }
+      }
+    }
+
+    const totalRecords = filteredRecords.length
+    const totalPages = totalRecords > 0 ? Math.ceil(totalRecords / pageSizeNumber) : 0
+    const safePage = totalPages > 0 ? Math.min(pageNumber, totalPages) : 1
+    const startIndex = (safePage - 1) * pageSizeNumber
+    const pageRecords =
+      totalRecords === 0 ? [] : filteredRecords.slice(startIndex, startIndex + pageSizeNumber)
+
+    const enrichedRecords = []
+    for (const record of pageRecords) {
+      const usage = toUsageObject(record)
+      const costData = CostCalculator.calculateCost(usage, record.model || 'unknown')
+      const computedCost =
+        typeof record.cost === 'number' ? record.cost : costData?.costs?.total || 0
+      const totalTokens =
+        record.totalTokens ||
+        usage.input_tokens +
+          usage.output_tokens +
+          usage.cache_creation_input_tokens +
+          usage.cache_read_input_tokens
+
+      const accountInfo = await resolveAccountInfo(record.accountId, record.accountType)
+      const resolvedAccountType = accountInfo?.type || record.accountType || 'unknown'
+
+      enrichedRecords.push({
+        timestamp: record.timestamp,
+        model: record.model || 'unknown',
+        accountId: record.accountId || null,
+        accountName: accountInfo?.name || null,
+        accountStatus: accountInfo?.status ?? null,
+        accountType: resolvedAccountType,
+        accountTypeName: accountTypeNames[resolvedAccountType] || '未知渠道',
+        inputTokens: usage.input_tokens,
+        outputTokens: usage.output_tokens,
+        cacheCreateTokens: usage.cache_creation_input_tokens,
+        cacheReadTokens: usage.cache_read_input_tokens,
+        ephemeral5mTokens: record.ephemeral5mTokens || 0,
+        ephemeral1hTokens: record.ephemeral1hTokens || 0,
+        totalTokens,
+        isLongContextRequest: record.isLongContext || record.isLongContextRequest || false,
+        cost: Number(computedCost.toFixed(6)),
+        costFormatted:
+          record.costFormatted ||
+          costData?.formatted?.total ||
+          CostCalculator.formatCost(computedCost),
+        costBreakdown: record.costBreakdown || {
+          input: costData?.costs?.input || 0,
+          output: costData?.costs?.output || 0,
+          cacheCreate: costData?.costs?.cacheWrite || 0,
+          cacheRead: costData?.costs?.cacheRead || 0,
+          total: costData?.costs?.total || computedCost
+        },
+        responseTime: record.responseTime || null
+      })
+    }
+
+    const accountOptions = []
+    for (const option of accountOptionMap.values()) {
+      const types = Array.from(option.accountTypes || [])
+
+      // 优先按历史出现的 accountType 解析，若失败则回退全量解析
+      let resolvedInfo = null
+      for (const type of types) {
+        resolvedInfo = await resolveAccountInfo(option.id, type)
+        if (resolvedInfo && resolvedInfo.name) {
+          break
+        }
+      }
+      if (!resolvedInfo) {
+        resolvedInfo = await resolveAccountInfo(option.id)
+      }
+
+      const chosenType = resolvedInfo?.type || types[0] || 'unknown'
+      const chosenTypeName = accountTypeNames[chosenType] || '未知渠道'
+
+      if (!resolvedInfo) {
+        logger.warn(`⚠️ 保留无法解析的账户筛选项: ${option.id}, types=${types.join(',') || 'none'}`)
+      }
+
+      accountOptions.push({
+        id: option.id,
+        name: resolvedInfo?.name || option.id,
+        accountType: chosenType,
+        accountTypeName: chosenTypeName,
+        rawTypes: types
+      })
+    }
+
+    return res.json({
+      success: true,
+      data: {
+        records: enrichedRecords,
+        pagination: {
+          currentPage: safePage,
+          pageSize: pageSizeNumber,
+          totalRecords,
+          totalPages,
+          hasNextPage: totalPages > 0 && safePage < totalPages,
+          hasPreviousPage: totalPages > 0 && safePage > 1
+        },
+        filters: {
+          startDate: startTime ? startTime.toISOString() : null,
+          endDate: endTime ? endTime.toISOString() : null,
+          model: model || null,
+          accountId: accountId || null,
+          sortOrder: normalizedSortOrder
+        },
+        apiKeyInfo: {
+          id: keyId,
+          name: apiKeyInfo.name || apiKeyInfo.label || keyId
+        },
+        summary: {
+          ...summary,
+          totalCost: Number(summary.totalCost.toFixed(6)),
+          avgCost:
+            summary.totalRequests > 0
+              ? Number((summary.totalCost / summary.totalRequests).toFixed(6))
+              : 0
+        },
+        availableFilters: {
+          models: Array.from(modelSet),
+          accounts: accountOptions,
+          dateRange: {
+            earliest: earliestTimestamp ? earliestTimestamp.toISOString() : null,
+            latest: latestTimestamp ? latestTimestamp.toISOString() : null
+          }
+        }
+      }
+    })
+  } catch (error) {
+    logger.error('❌ Failed to get API key usage records:', error)
+    return res
+      .status(500)
+      .json({ error: 'Failed to get API key usage records', message: error.message })
+  }
+})
+
+// 获取账户的请求记录时间线
+router.get('/accounts/:accountId/usage-records', authenticateAdmin, async (req, res) => {
+  try {
+    const { accountId } = req.params
+    const {
+      platform,
+      page = 1,
+      pageSize = 50,
+      startDate,
+      endDate,
+      model,
+      apiKeyId,
+      sortOrder = 'desc'
+    } = req.query
+
+    const pageNumber = Math.max(parseInt(page, 10) || 1, 1)
+    const pageSizeNumber = Math.min(Math.max(parseInt(pageSize, 10) || 50, 1), 200)
+    const normalizedSortOrder = sortOrder === 'asc' ? 'asc' : 'desc'
+
+    const startTime = startDate ? new Date(startDate) : null
+    const endTime = endDate ? new Date(endDate) : null
+
+    if (
+      (startDate && Number.isNaN(startTime?.getTime())) ||
+      (endDate && Number.isNaN(endTime?.getTime()))
+    ) {
+      return res.status(400).json({ success: false, error: 'Invalid date range' })
+    }
+
+    if (startTime && endTime && startTime > endTime) {
+      return res
+        .status(400)
+        .json({ success: false, error: 'Start date must be before or equal to end date' })
+    }
+
+    const accountInfo = await resolveAccountByPlatform(accountId, platform)
+    if (!accountInfo) {
+      return res.status(404).json({ success: false, error: 'Account not found' })
+    }
+
+    const allApiKeys = await apiKeyService.getAllApiKeys(true)
+    const apiKeyNameCache = new Map(
+      allApiKeys.map((key) => [key.id, key.name || key.label || key.id])
+    )
+
+    let keysToUse = apiKeyId ? allApiKeys.filter((key) => key.id === apiKeyId) : allApiKeys
+    if (apiKeyId && keysToUse.length === 0) {
+      keysToUse = [{ id: apiKeyId }]
+    }
+
+    const toUsageObject = (record) => ({
+      input_tokens: record.inputTokens || 0,
+      output_tokens: record.outputTokens || 0,
+      cache_creation_input_tokens: record.cacheCreateTokens || 0,
+      cache_read_input_tokens: record.cacheReadTokens || 0,
+      cache_creation: record.cacheCreation || record.cache_creation || null
+    })
+
+    const withinRange = (record) => {
+      if (!record.timestamp) {
+        return false
+      }
+      const ts = new Date(record.timestamp)
+      if (Number.isNaN(ts.getTime())) {
+        return false
+      }
+      if (startTime && ts < startTime) {
+        return false
+      }
+      if (endTime && ts > endTime) {
+        return false
+      }
+      return true
+    }
+
+    const filteredRecords = []
+    const modelSet = new Set()
+    const apiKeyOptionMap = new Map()
+    let earliestTimestamp = null
+    let latestTimestamp = null
+
+    const batchSize = 10
+    for (let i = 0; i < keysToUse.length; i += batchSize) {
+      const batch = keysToUse.slice(i, i + batchSize)
+      const batchResults = await Promise.all(
+        batch.map(async (key) => {
+          try {
+            const records = await redis.getUsageRecords(key.id, 5000)
+            return { keyId: key.id, records: records || [] }
+          } catch (error) {
+            logger.debug(`⚠️ Failed to get usage records for key ${key.id}: ${error.message}`)
+            return { keyId: key.id, records: [] }
+          }
+        })
+      )
+
+      for (const { keyId, records } of batchResults) {
+        const apiKeyName = apiKeyNameCache.get(keyId) || (await getApiKeyName(keyId))
+        for (const record of records) {
+          if (record.accountId !== accountId) {
+            continue
+          }
+          if (!withinRange(record)) {
+            continue
+          }
+          if (model && record.model !== model) {
+            continue
+          }
+
+          const accountType = record.accountType || accountInfo.platform || 'unknown'
+          const normalizedModel = record.model || 'unknown'
+
+          modelSet.add(normalizedModel)
+          apiKeyOptionMap.set(keyId, { id: keyId, name: apiKeyName })
+
+          if (record.timestamp) {
+            const ts = new Date(record.timestamp)
+            if (!Number.isNaN(ts.getTime())) {
+              if (!earliestTimestamp || ts < earliestTimestamp) {
+                earliestTimestamp = ts
+              }
+              if (!latestTimestamp || ts > latestTimestamp) {
+                latestTimestamp = ts
+              }
+            }
+          }
+
+          filteredRecords.push({
+            ...record,
+            model: normalizedModel,
+            accountType,
+            apiKeyId: keyId,
+            apiKeyName
+          })
+        }
+      }
+    }
+
+    filteredRecords.sort((a, b) => {
+      const aTime = new Date(a.timestamp).getTime()
+      const bTime = new Date(b.timestamp).getTime()
+      if (Number.isNaN(aTime) || Number.isNaN(bTime)) {
+        return 0
+      }
+      return normalizedSortOrder === 'asc' ? aTime - bTime : bTime - aTime
+    })
+
+    const summary = {
+      totalRequests: 0,
+      inputTokens: 0,
+      outputTokens: 0,
+      cacheCreateTokens: 0,
+      cacheReadTokens: 0,
+      totalTokens: 0,
+      totalCost: 0
+    }
+
+    for (const record of filteredRecords) {
+      const usage = toUsageObject(record)
+      const costData = CostCalculator.calculateCost(usage, record.model || 'unknown')
+      const computedCost =
+        typeof record.cost === 'number' ? record.cost : costData?.costs?.total || 0
+      const totalTokens =
+        record.totalTokens ||
+        usage.input_tokens +
+          usage.output_tokens +
+          usage.cache_creation_input_tokens +
+          usage.cache_read_input_tokens
+
+      summary.totalRequests += 1
+      summary.inputTokens += usage.input_tokens
+      summary.outputTokens += usage.output_tokens
+      summary.cacheCreateTokens += usage.cache_creation_input_tokens
+      summary.cacheReadTokens += usage.cache_read_input_tokens
+      summary.totalTokens += totalTokens
+      summary.totalCost += computedCost
+    }
+
+    const totalRecords = filteredRecords.length
+    const totalPages = totalRecords > 0 ? Math.ceil(totalRecords / pageSizeNumber) : 0
+    const safePage = totalPages > 0 ? Math.min(pageNumber, totalPages) : 1
+    const startIndex = (safePage - 1) * pageSizeNumber
+    const pageRecords =
+      totalRecords === 0 ? [] : filteredRecords.slice(startIndex, startIndex + pageSizeNumber)
+
+    const enrichedRecords = []
+    for (const record of pageRecords) {
+      const usage = toUsageObject(record)
+      const costData = CostCalculator.calculateCost(usage, record.model || 'unknown')
+      const computedCost =
+        typeof record.cost === 'number' ? record.cost : costData?.costs?.total || 0
+      const totalTokens =
+        record.totalTokens ||
+        usage.input_tokens +
+          usage.output_tokens +
+          usage.cache_creation_input_tokens +
+          usage.cache_read_input_tokens
+
+      enrichedRecords.push({
+        timestamp: record.timestamp,
+        model: record.model || 'unknown',
+        apiKeyId: record.apiKeyId,
+        apiKeyName: record.apiKeyName,
+        accountId,
+        accountName: accountInfo.name || accountInfo.email || accountId,
+        accountType: record.accountType,
+        accountTypeName: accountTypeNames[record.accountType] || '未知渠道',
+        inputTokens: usage.input_tokens,
+        outputTokens: usage.output_tokens,
+        cacheCreateTokens: usage.cache_creation_input_tokens,
+        cacheReadTokens: usage.cache_read_input_tokens,
+        ephemeral5mTokens: record.ephemeral5mTokens || 0,
+        ephemeral1hTokens: record.ephemeral1hTokens || 0,
+        totalTokens,
+        isLongContextRequest: record.isLongContext || record.isLongContextRequest || false,
+        cost: Number(computedCost.toFixed(6)),
+        costFormatted:
+          record.costFormatted ||
+          costData?.formatted?.total ||
+          CostCalculator.formatCost(computedCost),
+        costBreakdown: record.costBreakdown || {
+          input: costData?.costs?.input || 0,
+          output: costData?.costs?.output || 0,
+          cacheCreate: costData?.costs?.cacheWrite || 0,
+          cacheRead: costData?.costs?.cacheRead || 0,
+          total: costData?.costs?.total || computedCost
+        },
+        responseTime: record.responseTime || null
+      })
+    }
+
+    return res.json({
+      success: true,
+      data: {
+        records: enrichedRecords,
+        pagination: {
+          currentPage: safePage,
+          pageSize: pageSizeNumber,
+          totalRecords,
+          totalPages,
+          hasNextPage: totalPages > 0 && safePage < totalPages,
+          hasPreviousPage: totalPages > 0 && safePage > 1
+        },
+        filters: {
+          startDate: startTime ? startTime.toISOString() : null,
+          endDate: endTime ? endTime.toISOString() : null,
+          model: model || null,
+          apiKeyId: apiKeyId || null,
+          platform: accountInfo.platform,
+          sortOrder: normalizedSortOrder
+        },
+        accountInfo: {
+          id: accountId,
+          name: accountInfo.name || accountInfo.email || accountId,
+          platform: accountInfo.platform || platform || 'unknown',
+          status: accountInfo.status ?? accountInfo.isActive ?? null
+        },
+        summary: {
+          ...summary,
+          totalCost: Number(summary.totalCost.toFixed(6)),
+          avgCost:
+            summary.totalRequests > 0
+              ? Number((summary.totalCost / summary.totalRequests).toFixed(6))
+              : 0
+        },
+        availableFilters: {
+          models: Array.from(modelSet),
+          apiKeys: Array.from(apiKeyOptionMap.values()),
+          dateRange: {
+            earliest: earliestTimestamp ? earliestTimestamp.toISOString() : null,
+            latest: latestTimestamp ? latestTimestamp.toISOString() : null
+          }
+        }
+      }
+    })
+  } catch (error) {
+    logger.error('❌ Failed to get account usage records:', error)
+    return res
+      .status(500)
+      .json({ error: 'Failed to get account usage records', message: error.message })
+  }
+})
+
 module.exports = router

src/routes/admin/utils.js

@@ -33,7 +33,9 @@ function mapExpiryField(updates, accountType, accountId) {
   if ('expiresAt' in mappedUpdates) {
     mappedUpdates.subscriptionExpiresAt = mappedUpdates.expiresAt
     delete mappedUpdates.expiresAt
-    logger.info(`Mapping expiresAt to subscriptionExpiresAt for ${accountType} account ${accountId}`)
+    logger.info(
+      `Mapping expiresAt to subscriptionExpiresAt for ${accountType} account ${accountId}`
+    )
   }
   return mappedUpdates
 }

src/routes/api.js

@@ -11,7 +11,20 @@ const logger = require('../utils/logger')
 const { getEffectiveModel, parseVendorPrefixedModel } = require('../utils/modelHelper')
 const sessionHelper = require('../utils/sessionHelper')
 const { updateRateLimitCounters } = require('../utils/rateLimitHelper')
+const claudeRelayConfigService = require('../services/claudeRelayConfigService')
+const claudeAccountService = require('../services/claudeAccountService')
+const claudeConsoleAccountService = require('../services/claudeConsoleAccountService')
+const {
+  isWarmupRequest,
+  buildMockWarmupResponse,
+  sendMockWarmupStream
+} = require('../utils/warmupInterceptor')
 const { sanitizeUpstreamError } = require('../utils/errorSanitizer')
+const { dumpAnthropicMessagesRequest } = require('../utils/anthropicRequestDump')
+const {
+  handleAnthropicMessagesToGemini,
+  handleAnthropicCountTokensToGemini
+} = require('../services/anthropicGeminiBridgeService')
 const router = express.Router()
 
 function queueRateLimitUpdate(rateLimitInfo, usageSummary, model, context = '') {
@@ -37,21 +50,90 @@ function queueRateLimitUpdate(rateLimitInfo, usageSummary, model, context = '')
     })
 }
 
+/**
+ * 判断是否为旧会话（污染的会话）
+ * Claude Code 发送的请求特点：
+ * - messages 数组通常只有 1 个元素
+ * - 历史对话记录嵌套在单个 message 的 content 数组中
+ * - content 数组中包含 <system-reminder> 开头的系统注入内容
+ *
+ * 污染会话的特征：
+ * 1. messages.length > 1
+ * 2. messages.length === 1 但 content 中有多个用户输入
+ * 3. "warmup" 请求：单条简单消息 + 无 tools（真正新会话会带 tools）
+ *
+ * @param {Object} body - 请求体
+ * @returns {boolean} 是否为旧会话
+ */
+function isOldSession(body) {
+  const messages = body?.messages
+  const tools = body?.tools
+
+  if (!messages || messages.length === 0) {
+    return false
+  }
+
+  // 1. 多条消息 = 旧会话
+  if (messages.length > 1) {
+    return true
+  }
+
+  // 2. 单条消息，分析 content
+  const firstMessage = messages[0]
+  const content = firstMessage?.content
+
+  if (!content) {
+    return false
+  }
+
+  // 如果 content 是字符串，只有一条输入，需要检查 tools
+  if (typeof content === 'string') {
+    // 有 tools = 正常新会话，无 tools = 可疑
+    return !tools || tools.length === 0
+  }
+
+  // 如果 content 是数组，统计非 system-reminder 的元素
+  if (Array.isArray(content)) {
+    const userInputs = content.filter((item) => {
+      if (item.type !== 'text') {
+        return false
+      }
+      const text = item.text || ''
+      // 剔除以 <system-reminder> 开头的
+      return !text.trimStart().startsWith('<system-reminder>')
+    })
+
+    // 多个用户输入 = 旧会话
+    if (userInputs.length > 1) {
+      return true
+    }
+
+    // Warmup 检测：单个消息 + 无 tools = 旧会话
+    if (userInputs.length === 1 && (!tools || tools.length === 0)) {
+      return true
+    }
+  }
+
+  return false
+}
+
 // 🔧 共享的消息处理函数
 async function handleMessagesRequest(req, res) {
   try {
     const startTime = Date.now()
 
-    // Claude 服务权限校验，阻止未授权的 Key
-    if (
-      req.apiKey.permissions &&
-      req.apiKey.permissions !== 'all' &&
-      req.apiKey.permissions !== 'claude'
-    ) {
+    const forcedVendor = req._anthropicVendor || null
+    const requiredService =
+      forcedVendor === 'gemini-cli' || forcedVendor === 'antigravity' ? 'gemini' : 'claude'
+
+    if (!apiKeyService.hasPermission(req.apiKey?.permissions, requiredService)) {
       return res.status(403).json({
         error: {
           type: 'permission_error',
-          message: '此 API Key 无权访问 Claude 服务'
+          message:
+            requiredService === 'gemini'
+              ? '此 API Key 无权访问 Gemini 服务'
+              : '此 API Key 无权访问 Claude 服务'
         }
       })
     }
@@ -100,6 +182,25 @@ async function handleMessagesRequest(req, res) {
       }
     }
 
+    logger.api('📥 /v1/messages request received', {
+      model: req.body.model || null,
+      forcedVendor,
+      stream: req.body.stream === true
+    })
+
+    dumpAnthropicMessagesRequest(req, {
+      route: '/v1/messages',
+      forcedVendor,
+      model: req.body?.model || null,
+      stream: req.body?.stream === true
+    })
+
+    // /v1/messages 的扩展：按路径强制分流到 Gemini OAuth 账户（避免 model 前缀混乱）
+    if (forcedVendor === 'gemini-cli' || forcedVendor === 'antigravity') {
+      const baseModel = (req.body.model || '').trim()
+      return await handleAnthropicMessagesToGemini(req, res, { vendor: forcedVendor, baseModel })
+    }
+
     // 检查是否为流式请求
     const isStream = req.body.stream === true
 
@@ -122,12 +223,42 @@ async function handleMessagesRequest(req, res) {
     )
 
     if (isStream) {
+      // 🔍 检查客户端连接是否仍然有效（可能在并发排队等待期间断开）
+      if (res.destroyed || res.socket?.destroyed || res.writableEnded) {
+        logger.warn(
+          `⚠️ Client disconnected before stream response could start for key: ${req.apiKey?.name || 'unknown'}`
+        )
+        return undefined
+      }
+
       // 流式响应 - 只使用官方真实usage数据
       res.setHeader('Content-Type', 'text/event-stream')
       res.setHeader('Cache-Control', 'no-cache')
       res.setHeader('Connection', 'keep-alive')
       res.setHeader('Access-Control-Allow-Origin', '*')
       res.setHeader('X-Accel-Buffering', 'no') // 禁用 Nginx 缓冲
+      // ⚠️ 检查 headers 是否已发送（可能在排队心跳时已设置）
+      if (!res.headersSent) {
+        res.setHeader('Content-Type', 'text/event-stream')
+        res.setHeader('Cache-Control', 'no-cache')
+        // ⚠️ 关键修复：尊重 auth.js 提前设置的 Connection: close
+        // 当并发队列功能启用时，auth.js 会设置 Connection: close 来禁用 Keep-Alive
+        // 这里只在没有设置过 Connection 头时才设置 keep-alive
+        const existingConnection = res.getHeader('Connection')
+        if (!existingConnection) {
+          res.setHeader('Connection', 'keep-alive')
+        } else {
+          logger.api(
+            `🔌 [STREAM] Preserving existing Connection header: ${existingConnection} for key: ${req.apiKey?.name || 'unknown'}`
+          )
+        }
+        res.setHeader('Access-Control-Allow-Origin', '*')
+        res.setHeader('X-Accel-Buffering', 'no') // 禁用 Nginx 缓冲
+      } else {
+        logger.debug(
+          `📤 [STREAM] Headers already sent, skipping setHeader for key: ${req.apiKey?.name || 'unknown'}`
+        )
+      }
 
       // 禁用 Nagle 算法，确保数据立即发送
       if (res.socket && typeof res.socket.setNoDelay === 'function') {
@@ -141,6 +272,56 @@ async function handleMessagesRequest(req, res) {
       // 生成会话哈希用于sticky会话
       const sessionHash = sessionHelper.generateSessionHash(req.body)
 
+      // 🔒 全局会话绑定验证
+      let forcedAccount = null
+      let needSessionBinding = false
+      let originalSessionIdForBinding = null
+
+      try {
+        const globalBindingEnabled = await claudeRelayConfigService.isGlobalSessionBindingEnabled()
+
+        if (globalBindingEnabled) {
+          const originalSessionId = claudeRelayConfigService.extractOriginalSessionId(req.body)
+
+          if (originalSessionId) {
+            const validation = await claudeRelayConfigService.validateNewSession(
+              req.body,
+              originalSessionId
+            )
+
+            if (!validation.valid) {
+              logger.api(
+                `❌ Session binding validation failed: ${validation.code} for session ${originalSessionId}`
+              )
+              return res.status(403).json({
+                error: {
+                  type: 'session_binding_error',
+                  message: validation.error
+                }
+              })
+            }
+
+            // 如果已有绑定，使用绑定的账户
+            if (validation.binding) {
+              forcedAccount = validation.binding
+              logger.api(
+                `🔗 Using bound account for session ${originalSessionId}: ${forcedAccount.accountId}`
+              )
+            }
+
+            // 标记需要在调度成功后建立绑定
+            if (validation.isNewSession) {
+              needSessionBinding = true
+              originalSessionIdForBinding = originalSessionId
+              logger.api(`📝 New session detected, will create binding: ${originalSessionId}`)
+            }
+          }
+        }
+      } catch (error) {
+        logger.error('❌ Error in global session binding check:', error)
+        // 配置服务出错时不阻断请求
+      }
+
       // 使用统一调度选择账号（传递请求的模型）
       const requestedModel = req.body.model
       let accountId
@@ -149,10 +330,21 @@ async function handleMessagesRequest(req, res) {
         const selection = await unifiedClaudeScheduler.selectAccountForApiKey(
           req.apiKey,
           sessionHash,
-          requestedModel
+          requestedModel,
+          forcedAccount
         )
         ;({ accountId, accountType } = selection)
       } catch (error) {
+        // 处理会话绑定账户不可用的错误
+        if (error.code === 'SESSION_BINDING_ACCOUNT_UNAVAILABLE') {
+          const errorMessage = await claudeRelayConfigService.getSessionBindingErrorMessage()
+          return res.status(403).json({
+            error: {
+              type: 'session_binding_error',
+              message: errorMessage
+            }
+          })
+        }
         if (error.code === 'CLAUDE_DEDICATED_RATE_LIMITED') {
           const limitMessage = claudeRelayService._buildStandardRateLimitMessage(
             error.rateLimitEndAt
@@ -170,14 +362,72 @@ async function handleMessagesRequest(req, res) {
         throw error
       }
 
+      // 🔗 在成功调度后建立会话绑定（仅 claude-official 类型）
+      // claude-official 只接受：1) 新会话 2) 已绑定的会话
+      if (
+        needSessionBinding &&
+        originalSessionIdForBinding &&
+        accountId &&
+        accountType === 'claude-official'
+      ) {
+        // 🚫 检测旧会话（污染的会话）
+        if (isOldSession(req.body)) {
+          const cfg = await claudeRelayConfigService.getConfig()
+          logger.warn(
+            `🚫 Old session rejected: sessionId=${originalSessionIdForBinding}, messages.length=${req.body?.messages?.length}, tools.length=${req.body?.tools?.length || 0}, isOldSession=true`
+          )
+          return res.status(400).json({
+            error: {
+              type: 'session_binding_error',
+              message: cfg.sessionBindingErrorMessage || '你的本地session已污染，请清理后使用。'
+            }
+          })
+        }
+
+        // 创建绑定
+        try {
+          await claudeRelayConfigService.setOriginalSessionBinding(
+            originalSessionIdForBinding,
+            accountId,
+            accountType
+          )
+        } catch (bindingError) {
+          logger.warn(`⚠️ Failed to create session binding:`, bindingError)
+        }
+      }
+
+      // 🔥 预热请求拦截检查（在转发之前）
+      if (accountType === 'claude-official' || accountType === 'claude-console') {
+        const account =
+          accountType === 'claude-official'
+            ? await claudeAccountService.getAccount(accountId)
+            : await claudeConsoleAccountService.getAccount(accountId)
+
+        if (account?.interceptWarmup === 'true' && isWarmupRequest(req.body)) {
+          logger.api(`🔥 Warmup request intercepted for account: ${account.name} (${accountId})`)
+          if (isStream) {
+            return sendMockWarmupStream(res, req.body.model)
+          } else {
+            return res.json(buildMockWarmupResponse(req.body.model))
+          }
+        }
+      }
+
       // 根据账号类型选择对应的转发服务并调用
       if (accountType === 'claude-official') {
         // 官方Claude账号使用原有的转发服务（会自己选择账号）
+        // 🧹 内存优化：提取需要的值，避免闭包捕获整个 req 对象
+        const _apiKeyId = req.apiKey.id
+        const _rateLimitInfo = req.rateLimitInfo
+        const _requestBody = req.body // 传递后清除引用
+        const _apiKey = req.apiKey
+        const _headers = req.headers
+
         await claudeRelayService.relayStreamRequestWithUsageCapture(
-          req.body,
-          req.apiKey,
+          _requestBody,
+          _apiKey,
           res,
-          req.headers,
+          _headers,
           (usageData) => {
             // 回调函数：当检测到完整usage数据时记录真实token使用量
             logger.info(
@@ -227,13 +477,13 @@ async function handleMessagesRequest(req, res) {
               }
 
               apiKeyService
-                .recordUsageWithDetails(req.apiKey.id, usageObject, model, usageAccountId, 'claude')
+                .recordUsageWithDetails(_apiKeyId, usageObject, model, usageAccountId, 'claude')
                 .catch((error) => {
                   logger.error('❌ Failed to record stream usage:', error)
                 })
 
               queueRateLimitUpdate(
-                req.rateLimitInfo,
+                _rateLimitInfo,
                 {
                   inputTokens,
                   outputTokens,
@@ -258,11 +508,18 @@ async function handleMessagesRequest(req, res) {
         )
       } else if (accountType === 'claude-console') {
         // Claude Console账号使用Console转发服务（需要传递accountId）
+        // 🧹 内存优化：提取需要的值
+        const _apiKeyIdConsole = req.apiKey.id
+        const _rateLimitInfoConsole = req.rateLimitInfo
+        const _requestBodyConsole = req.body
+        const _apiKeyConsole = req.apiKey
+        const _headersConsole = req.headers
+
         await claudeConsoleRelayService.relayStreamRequestWithUsageCapture(
-          req.body,
-          req.apiKey,
+          _requestBodyConsole,
+          _apiKeyConsole,
           res,
-          req.headers,
+          _headersConsole,
           (usageData) => {
             // 回调函数：当检测到完整usage数据时记录真实token使用量
             logger.info(
@@ -313,7 +570,7 @@ async function handleMessagesRequest(req, res) {
 
               apiKeyService
                 .recordUsageWithDetails(
-                  req.apiKey.id,
+                  _apiKeyIdConsole,
                   usageObject,
                   model,
                   usageAccountId,
@@ -324,7 +581,7 @@ async function handleMessagesRequest(req, res) {
                 })
 
               queueRateLimitUpdate(
-                req.rateLimitInfo,
+                _rateLimitInfoConsole,
                 {
                   inputTokens,
                   outputTokens,
@@ -350,6 +607,11 @@ async function handleMessagesRequest(req, res) {
         )
       } else if (accountType === 'bedrock') {
         // Bedrock账号使用Bedrock转发服务
+        // 🧹 内存优化：提取需要的值
+        const _apiKeyIdBedrock = req.apiKey.id
+        const _rateLimitInfoBedrock = req.rateLimitInfo
+        const _requestBodyBedrock = req.body
+
         try {
           const bedrockAccountResult = await bedrockAccountService.getAccount(accountId)
           if (!bedrockAccountResult.success) {
@@ -357,7 +619,7 @@ async function handleMessagesRequest(req, res) {
           }
 
           const result = await bedrockRelayService.handleStreamRequest(
-            req.body,
+            _requestBodyBedrock,
             bedrockAccountResult.data,
             res
           )
@@ -368,13 +630,21 @@ async function handleMessagesRequest(req, res) {
             const outputTokens = result.usage.output_tokens || 0
 
             apiKeyService
-              .recordUsage(req.apiKey.id, inputTokens, outputTokens, 0, 0, result.model, accountId)
+              .recordUsage(
+                _apiKeyIdBedrock,
+                inputTokens,
+                outputTokens,
+                0,
+                0,
+                result.model,
+                accountId
+              )
               .catch((error) => {
                 logger.error('❌ Failed to record Bedrock stream usage:', error)
               })
 
             queueRateLimitUpdate(
-              req.rateLimitInfo,
+              _rateLimitInfoBedrock,
               {
                 inputTokens,
                 outputTokens,
@@ -399,11 +669,18 @@ async function handleMessagesRequest(req, res) {
         }
       } else if (accountType === 'ccr') {
         // CCR账号使用CCR转发服务（需要传递accountId）
+        // 🧹 内存优化：提取需要的值
+        const _apiKeyIdCcr = req.apiKey.id
+        const _rateLimitInfoCcr = req.rateLimitInfo
+        const _requestBodyCcr = req.body
+        const _apiKeyCcr = req.apiKey
+        const _headersCcr = req.headers
+
         await ccrRelayService.relayStreamRequestWithUsageCapture(
-          req.body,
-          req.apiKey,
+          _requestBodyCcr,
+          _apiKeyCcr,
           res,
-          req.headers,
+          _headersCcr,
           (usageData) => {
             // 回调函数：当检测到完整usage数据时记录真实token使用量
             logger.info(
@@ -453,13 +730,13 @@ async function handleMessagesRequest(req, res) {
               }
 
               apiKeyService
-                .recordUsageWithDetails(req.apiKey.id, usageObject, model, usageAccountId, 'ccr')
+                .recordUsageWithDetails(_apiKeyIdCcr, usageObject, model, usageAccountId, 'ccr')
                 .catch((error) => {
                   logger.error('❌ Failed to record CCR stream usage:', error)
                 })
 
               queueRateLimitUpdate(
-                req.rateLimitInfo,
+                _rateLimitInfoCcr,
                 {
                   inputTokens,
                   outputTokens,
@@ -494,15 +771,121 @@ async function handleMessagesRequest(req, res) {
         }
       }, 1000) // 1秒后检查
     } else {
+      // 🧹 内存优化：提取需要的值，避免后续回调捕获整个 req
+      const _apiKeyIdNonStream = req.apiKey.id
+      const _apiKeyNameNonStream = req.apiKey.name
+      const _rateLimitInfoNonStream = req.rateLimitInfo
+      const _requestBodyNonStream = req.body
+      const _apiKeyNonStream = req.apiKey
+      const _headersNonStream = req.headers
+
+      // 🔍 检查客户端连接是否仍然有效（可能在并发排队等待期间断开）
+      if (res.destroyed || res.socket?.destroyed || res.writableEnded) {
+        logger.warn(
+          `⚠️ Client disconnected before non-stream request could start for key: ${_apiKeyNameNonStream || 'unknown'}`
+        )
+        return undefined
+      }
+
       // 非流式响应 - 只使用官方真实usage数据
       logger.info('📄 Starting non-streaming request', {
-        apiKeyId: req.apiKey.id,
-        apiKeyName: req.apiKey.name
+        apiKeyId: _apiKeyIdNonStream,
+        apiKeyName: _apiKeyNameNonStream
       })
 
+      // 📊 监听 socket 事件以追踪连接状态变化
+      const nonStreamSocket = res.socket
+      let _clientClosedConnection = false
+      let _socketCloseTime = null
+
+      if (nonStreamSocket) {
+        const onSocketEnd = () => {
+          _clientClosedConnection = true
+          _socketCloseTime = Date.now()
+          logger.warn(
+            `⚠️ [NON-STREAM] Socket 'end' event - client sent FIN | key: ${req.apiKey?.name}, ` +
+              `requestId: ${req.requestId}, elapsed: ${Date.now() - startTime}ms`
+          )
+        }
+        const onSocketClose = () => {
+          _clientClosedConnection = true
+          logger.warn(
+            `⚠️ [NON-STREAM] Socket 'close' event | key: ${req.apiKey?.name}, ` +
+              `requestId: ${req.requestId}, elapsed: ${Date.now() - startTime}ms, ` +
+              `hadError: ${nonStreamSocket.destroyed}`
+          )
+        }
+        const onSocketError = (err) => {
+          logger.error(
+            `❌ [NON-STREAM] Socket error | key: ${req.apiKey?.name}, ` +
+              `requestId: ${req.requestId}, error: ${err.message}`
+          )
+        }
+
+        nonStreamSocket.once('end', onSocketEnd)
+        nonStreamSocket.once('close', onSocketClose)
+        nonStreamSocket.once('error', onSocketError)
+
+        // 清理监听器（在响应结束后）
+        res.once('finish', () => {
+          nonStreamSocket.removeListener('end', onSocketEnd)
+          nonStreamSocket.removeListener('close', onSocketClose)
+          nonStreamSocket.removeListener('error', onSocketError)
+        })
+      }
+
       // 生成会话哈希用于sticky会话
       const sessionHash = sessionHelper.generateSessionHash(req.body)
 
+      // 🔒 全局会话绑定验证（非流式）
+      let forcedAccountNonStream = null
+      let needSessionBindingNonStream = false
+      let originalSessionIdForBindingNonStream = null
+
+      try {
+        const globalBindingEnabled = await claudeRelayConfigService.isGlobalSessionBindingEnabled()
+
+        if (globalBindingEnabled) {
+          const originalSessionId = claudeRelayConfigService.extractOriginalSessionId(req.body)
+
+          if (originalSessionId) {
+            const validation = await claudeRelayConfigService.validateNewSession(
+              req.body,
+              originalSessionId
+            )
+
+            if (!validation.valid) {
+              logger.api(
+                `❌ Session binding validation failed (non-stream): ${validation.code} for session ${originalSessionId}`
+              )
+              return res.status(403).json({
+                error: {
+                  type: 'session_binding_error',
+                  message: validation.error
+                }
+              })
+            }
+
+            if (validation.binding) {
+              forcedAccountNonStream = validation.binding
+              logger.api(
+                `🔗 Using bound account for session (non-stream) ${originalSessionId}: ${forcedAccountNonStream.accountId}`
+              )
+            }
+
+            if (validation.isNewSession) {
+              needSessionBindingNonStream = true
+              originalSessionIdForBindingNonStream = originalSessionId
+              logger.api(
+                `📝 New session detected (non-stream), will create binding: ${originalSessionId}`
+              )
+            }
+          }
+        }
+      } catch (error) {
+        logger.error('❌ Error in global session binding check (non-stream):', error)
+      }
+
       // 使用统一调度选择账号（传递请求的模型）
       const requestedModel = req.body.model
       let accountId
@@ -511,10 +894,20 @@ async function handleMessagesRequest(req, res) {
         const selection = await unifiedClaudeScheduler.selectAccountForApiKey(
           req.apiKey,
           sessionHash,
-          requestedModel
+          requestedModel,
+          forcedAccountNonStream
         )
         ;({ accountId, accountType } = selection)
       } catch (error) {
+        if (error.code === 'SESSION_BINDING_ACCOUNT_UNAVAILABLE') {
+          const errorMessage = await claudeRelayConfigService.getSessionBindingErrorMessage()
+          return res.status(403).json({
+            error: {
+              type: 'session_binding_error',
+              message: errorMessage
+            }
+          })
+        }
         if (error.code === 'CLAUDE_DEDICATED_RATE_LIMITED') {
           const limitMessage = claudeRelayService._buildStandardRateLimitMessage(
             error.rateLimitEndAt
@@ -527,6 +920,55 @@ async function handleMessagesRequest(req, res) {
         throw error
       }
 
+      // 🔗 在成功调度后建立会话绑定（非流式，仅 claude-official 类型）
+      // claude-official 只接受：1) 新会话 2) 已绑定的会话
+      if (
+        needSessionBindingNonStream &&
+        originalSessionIdForBindingNonStream &&
+        accountId &&
+        accountType === 'claude-official'
+      ) {
+        // 🚫 检测旧会话（污染的会话）
+        if (isOldSession(req.body)) {
+          const cfg = await claudeRelayConfigService.getConfig()
+          logger.warn(
+            `🚫 Old session rejected (non-stream): sessionId=${originalSessionIdForBindingNonStream}, messages.length=${req.body?.messages?.length}, tools.length=${req.body?.tools?.length || 0}, isOldSession=true`
+          )
+          return res.status(400).json({
+            error: {
+              type: 'session_binding_error',
+              message: cfg.sessionBindingErrorMessage || '你的本地session已污染，请清理后使用。'
+            }
+          })
+        }
+
+        // 创建绑定
+        try {
+          await claudeRelayConfigService.setOriginalSessionBinding(
+            originalSessionIdForBindingNonStream,
+            accountId,
+            accountType
+          )
+        } catch (bindingError) {
+          logger.warn(`⚠️ Failed to create session binding (non-stream):`, bindingError)
+        }
+      }
+
+      // 🔥 预热请求拦截检查（非流式，在转发之前）
+      if (accountType === 'claude-official' || accountType === 'claude-console') {
+        const account =
+          accountType === 'claude-official'
+            ? await claudeAccountService.getAccount(accountId)
+            : await claudeConsoleAccountService.getAccount(accountId)
+
+        if (account?.interceptWarmup === 'true' && isWarmupRequest(_requestBodyNonStream)) {
+          logger.api(
+            `🔥 Warmup request intercepted (non-stream) for account: ${account.name} (${accountId})`
+          )
+          return res.json(buildMockWarmupResponse(_requestBodyNonStream.model))
+        }
+      }
+
       // 根据账号类型选择对应的转发服务
       let response
       logger.debug(`[DEBUG] Request query params: ${JSON.stringify(req.query)}`)
@@ -536,11 +978,11 @@ async function handleMessagesRequest(req, res) {
       if (accountType === 'claude-official') {
         // 官方Claude账号使用原有的转发服务
         response = await claudeRelayService.relayRequest(
-          req.body,
-          req.apiKey,
-          req,
+          _requestBodyNonStream,
+          _apiKeyNonStream,
+          req, // clientRequest 用于断开检测，保留但服务层已优化
           res,
-          req.headers
+          _headersNonStream
         )
       } else if (accountType === 'claude-console') {
         // Claude Console账号使用Console转发服务
@@ -548,11 +990,11 @@ async function handleMessagesRequest(req, res) {
           `[DEBUG] Calling claudeConsoleRelayService.relayRequest with accountId: ${accountId}`
         )
         response = await claudeConsoleRelayService.relayRequest(
-          req.body,
-          req.apiKey,
-          req,
+          _requestBodyNonStream,
+          _apiKeyNonStream,
+          req, // clientRequest 保留用于断开检测
           res,
-          req.headers,
+          _headersNonStream,
           accountId
         )
       } else if (accountType === 'bedrock') {
@@ -564,9 +1006,9 @@ async function handleMessagesRequest(req, res) {
           }
 
           const result = await bedrockRelayService.handleNonStreamRequest(
-            req.body,
+            _requestBodyNonStream,
             bedrockAccountResult.data,
-            req.headers
+            _headersNonStream
           )
 
           // 构建标准响应格式
@@ -596,11 +1038,11 @@ async function handleMessagesRequest(req, res) {
         // CCR账号使用CCR转发服务
         logger.debug(`[DEBUG] Calling ccrRelayService.relayRequest with accountId: ${accountId}`)
         response = await ccrRelayService.relayRequest(
-          req.body,
-          req.apiKey,
-          req,
+          _requestBodyNonStream,
+          _apiKeyNonStream,
+          req, // clientRequest 保留用于断开检测
           res,
-          req.headers,
+          _headersNonStream,
           accountId
         )
       }
@@ -611,6 +1053,15 @@ async function handleMessagesRequest(req, res) {
         bodyLength: response.body ? response.body.length : 0
       })
 
+      // 🔍 检查客户端连接是否仍然有效
+      // 在长时间请求过程中，客户端可能已经断开连接（超时、用户取消等）
+      if (res.destroyed || res.socket?.destroyed || res.writableEnded) {
+        logger.warn(
+          `⚠️ Client disconnected before non-stream response could be sent for key: ${req.apiKey?.name || 'unknown'}`
+        )
+        return undefined
+      }
+
       res.status(response.statusCode)
 
       // 设置响应头，避免 Content-Length 和 Transfer-Encoding 冲突
@@ -640,14 +1091,14 @@ async function handleMessagesRequest(req, res) {
           const cacheCreateTokens = jsonData.usage.cache_creation_input_tokens || 0
           const cacheReadTokens = jsonData.usage.cache_read_input_tokens || 0
           // Parse the model to remove vendor prefix if present (e.g., "ccr,gemini-2.5-pro" -> "gemini-2.5-pro")
-          const rawModel = jsonData.model || req.body.model || 'unknown'
-          const { baseModel } = parseVendorPrefixedModel(rawModel)
-          const model = baseModel || rawModel
+          const rawModel = jsonData.model || _requestBodyNonStream.model || 'unknown'
+          const { baseModel: usageBaseModel } = parseVendorPrefixedModel(rawModel)
+          const model = usageBaseModel || rawModel
 
           // 记录真实的token使用量（包含模型信息和所有4种token以及账户ID）
           const { accountId: responseAccountId } = response
           await apiKeyService.recordUsage(
-            req.apiKey.id,
+            _apiKeyIdNonStream,
             inputTokens,
             outputTokens,
             cacheCreateTokens,
@@ -657,7 +1108,7 @@ async function handleMessagesRequest(req, res) {
           )
 
           await queueRateLimitUpdate(
-            req.rateLimitInfo,
+            _rateLimitInfoNonStream,
             {
               inputTokens,
               outputTokens,
@@ -676,10 +1127,12 @@ async function handleMessagesRequest(req, res) {
           logger.warn('⚠️ No usage data found in Claude API JSON response')
         }
 
+        // 使用 Express 内建的 res.json() 发送响应（简单可靠）
         res.json(jsonData)
       } catch (parseError) {
         logger.warn('⚠️ Failed to parse Claude API response as JSON:', parseError.message)
         logger.info('📄 Raw response body:', response.body)
+        // 使用 Express 内建的 res.send() 发送响应（简单可靠）
         res.send(response.body)
       }
 
@@ -816,6 +1269,65 @@ router.post('/claude/v1/messages', authenticateApiKey, handleMessagesRequest)
 // 📋 模型列表端点 - 支持 Claude, OpenAI, Gemini
 router.get('/v1/models', authenticateApiKey, async (req, res) => {
   try {
+    // Claude Code / Anthropic baseUrl 的分流：/antigravity/api/v1/models 返回 Antigravity 实时模型列表
+    //（通过 v1internal:fetchAvailableModels），避免依赖静态 modelService 列表。
+    const forcedVendor = req._anthropicVendor || null
+    if (forcedVendor === 'antigravity') {
+      if (!apiKeyService.hasPermission(req.apiKey?.permissions, 'gemini')) {
+        return res.status(403).json({
+          error: {
+            type: 'permission_error',
+            message: '此 API Key 无权访问 Gemini 服务'
+          }
+        })
+      }
+
+      const unifiedGeminiScheduler = require('../services/unifiedGeminiScheduler')
+      const geminiAccountService = require('../services/geminiAccountService')
+
+      let accountSelection
+      try {
+        accountSelection = await unifiedGeminiScheduler.selectAccountForApiKey(
+          req.apiKey,
+          null,
+          null,
+          { oauthProvider: 'antigravity' }
+        )
+      } catch (error) {
+        logger.error('Failed to select Gemini OAuth account (antigravity models):', error)
+        return res.status(503).json({ error: 'No available Gemini OAuth accounts' })
+      }
+
+      const account = await geminiAccountService.getAccount(accountSelection.accountId)
+      if (!account) {
+        return res.status(503).json({ error: 'Gemini OAuth account not found' })
+      }
+
+      let proxyConfig = null
+      if (account.proxy) {
+        try {
+          proxyConfig =
+            typeof account.proxy === 'string' ? JSON.parse(account.proxy) : account.proxy
+        } catch (e) {
+          logger.warn('Failed to parse proxy configuration:', e)
+        }
+      }
+
+      const models = await geminiAccountService.fetchAvailableModelsAntigravity(
+        account.accessToken,
+        proxyConfig,
+        account.refreshToken
+      )
+
+      // 可选：根据 API Key 的模型限制过滤（黑名单语义）
+      let filteredModels = models
+      if (req.apiKey.enableModelRestriction && req.apiKey.restrictedModels?.length > 0) {
+        filteredModels = models.filter((model) => !req.apiKey.restrictedModels.includes(model.id))
+      }
+
+      return res.json({ object: 'list', data: filteredModels })
+    }
+
     const modelService = require('../services/modelService')
 
     // 从 modelService 获取所有支持的模型
@@ -824,7 +1336,8 @@ router.get('/v1/models', authenticateApiKey, async (req, res) => {
     // 可选：根据 API Key 的模型限制过滤
     let filteredModels = models
     if (req.apiKey.enableModelRestriction && req.apiKey.restrictedModels?.length > 0) {
-      filteredModels = models.filter((model) => req.apiKey.restrictedModels.includes(model.id))
+      // 将 restrictedModels 视为黑名单：过滤掉受限模型
+      filteredModels = models.filter((model) => !req.apiKey.restrictedModels.includes(model.id))
     }
 
     res.json({
@@ -951,20 +1464,62 @@ router.get('/v1/organizations/:org_id/usage', authenticateApiKey, async (req, re
 
 // 🔢 Token计数端点 - count_tokens beta API
 router.post('/v1/messages/count_tokens', authenticateApiKey, async (req, res) => {
-  // 检查权限
-  if (
-    req.apiKey.permissions &&
-    req.apiKey.permissions !== 'all' &&
-    req.apiKey.permissions !== 'claude'
-  ) {
+  // 按路径强制分流到 Gemini OAuth 账户（避免 model 前缀混乱）
+  const forcedVendor = req._anthropicVendor || null
+  const requiredService =
+    forcedVendor === 'gemini-cli' || forcedVendor === 'antigravity' ? 'gemini' : 'claude'
+
+  if (!apiKeyService.hasPermission(req.apiKey?.permissions, requiredService)) {
     return res.status(403).json({
       error: {
         type: 'permission_error',
-        message: 'This API key does not have permission to access Claude'
+        message:
+          requiredService === 'gemini'
+            ? 'This API key does not have permission to access Gemini'
+            : 'This API key does not have permission to access Claude'
       }
     })
   }
 
+  if (requiredService === 'gemini') {
+    return await handleAnthropicCountTokensToGemini(req, res, { vendor: forcedVendor })
+  }
+
+  // 🔗 会话绑定验证（与 messages 端点保持一致）
+  const originalSessionId = claudeRelayConfigService.extractOriginalSessionId(req.body)
+  const sessionValidation = await claudeRelayConfigService.validateNewSession(
+    req.body,
+    originalSessionId
+  )
+
+  if (!sessionValidation.valid) {
+    logger.warn(
+      `🚫 Session binding validation failed (count_tokens): ${sessionValidation.code} for session ${originalSessionId}`
+    )
+    return res.status(400).json({
+      error: {
+        type: 'session_binding_error',
+        message: sessionValidation.error
+      }
+    })
+  }
+
+  // 🔗 检测旧会话（污染的会话）- 仅对需要绑定的新会话检查
+  if (sessionValidation.isNewSession && originalSessionId) {
+    if (isOldSession(req.body)) {
+      const cfg = await claudeRelayConfigService.getConfig()
+      logger.warn(
+        `🚫 Old session rejected (count_tokens): sessionId=${originalSessionId}, messages.length=${req.body?.messages?.length}, tools.length=${req.body?.tools?.length || 0}, isOldSession=true`
+      )
+      return res.status(400).json({
+        error: {
+          type: 'session_binding_error',
+          message: cfg.sessionBindingErrorMessage || '你的本地session已污染，请清理后使用。'
+        }
+      })
+    }
+  }
+
   logger.info(`🔢 Processing token count request for key: ${req.apiKey.name}`)
 
   const sessionHash = sessionHelper.generateSessionHash(req.body)
@@ -1003,6 +1558,17 @@ router.post('/v1/messages/count_tokens', authenticateApiKey, async (req, res) =>
       })
     }
 
+    // 🔍 claude-console 账户特殊处理：检查 count_tokens 端点是否可用
+    if (accountType === 'claude-console') {
+      const isUnavailable = await claudeConsoleAccountService.isCountTokensUnavailable(accountId)
+      if (isUnavailable) {
+        logger.info(
+          `⏭️ count_tokens unavailable for Claude Console account ${accountId}, returning fallback response`
+        )
+        return { fallbackResponse: true }
+      }
+    }
+
     const relayOptions = {
       skipUsageRecord: true,
       customPath: '/v1/messages/count_tokens'
@@ -1028,6 +1594,23 @@ router.post('/v1/messages/count_tokens', authenticateApiKey, async (req, res) =>
             relayOptions
           )
 
+    // 🔍 claude-console 账户：检测上游 404 响应并标记
+    if (accountType === 'claude-console' && response.statusCode === 404) {
+      logger.warn(
+        `⚠️ count_tokens endpoint returned 404 for Claude Console account ${accountId}, marking as unavailable`
+      )
+      // 标记失败不应影响 fallback 响应
+      try {
+        await claudeConsoleAccountService.markCountTokensUnavailable(accountId)
+      } catch (markError) {
+        logger.error(
+          `❌ Failed to mark count_tokens unavailable for account ${accountId}, but will still return fallback:`,
+          markError
+        )
+      }
+      return { fallbackResponse: true }
+    }
+
     res.status(response.statusCode)
 
     const skipHeaders = ['content-encoding', 'transfer-encoding', 'content-length']
@@ -1050,11 +1633,21 @@ router.post('/v1/messages/count_tokens', authenticateApiKey, async (req, res) =>
     }
 
     logger.info(`✅ Token count request completed for key: ${req.apiKey.name}`)
+    return { fallbackResponse: false }
   }
 
   while (attempt < maxAttempts) {
     try {
-      await processRequest()
+      const result = await processRequest()
+
+      // 🔍 处理 fallback 响应（claude-console 账户 count_tokens 不可用）
+      if (result && result.fallbackResponse) {
+        if (!res.headersSent) {
+          return res.status(200).json({ input_tokens: 0 })
+        }
+        return
+      }
+
       return
     } catch (error) {
       if (error.code === 'CONSOLE_ACCOUNT_CONCURRENCY_FULL') {
@@ -1129,5 +1722,10 @@ router.post('/v1/messages/count_tokens', authenticateApiKey, async (req, res) =>
   }
 })
 
+// Claude Code 客户端遥测端点 - 返回成功响应避免 404 日志
+router.post('/api/event_logging/batch', (req, res) => {
+  res.status(200).json({ success: true })
+})
+
 module.exports = router
 module.exports.handleMessagesRequest = handleMessagesRequest

src/routes/apiStats.js

@@ -5,6 +5,7 @@ const apiKeyService = require('../services/apiKeyService')
 const CostCalculator = require('../utils/costCalculator')
 const claudeAccountService = require('../services/claudeAccountService')
 const openaiAccountService = require('../services/openaiAccountService')
+const { createClaudeTestPayload } = require('../utils/testPayloadHelper')
 
 const router = express.Router()
 
@@ -154,7 +155,7 @@ router.post('/api/user-stats', async (req, res) => {
         restrictedModels,
         enableClientRestriction: keyData.enableClientRestriction === 'true',
         allowedClients,
-        permissions: keyData.permissions || 'all',
+        permissions: keyData.permissions,
         // 添加激活相关字段
         expirationMode: keyData.expirationMode || 'fixed',
         isActivated: keyData.isActivated === 'true',
@@ -205,74 +206,85 @@ router.post('/api/user-stats', async (req, res) => {
     // 获取验证结果中的完整keyData（包含isActive状态和cost信息）
     const fullKeyData = keyData
 
-    // 计算总费用 - 使用与模型统计相同的逻辑（按模型分别计算）
+    // 🔧 FIX: 使用 allTimeCost 而不是扫描月度键
+    // 计算总费用 - 优先使用持久化的总费用计数器
     let totalCost = 0
     let formattedCost = '$0.000000'
 
     try {
       const client = redis.getClientSafe()
 
-      // 获取所有月度模型统计（与model-stats接口相同的逻辑）
-      const allModelKeys = await client.keys(`usage:${keyId}:model:monthly:*:*`)
-      const modelUsageMap = new Map()
+      // 读取累积的总费用（没有 TTL 的持久键）
+      const totalCostKey = `usage:cost:total:${keyId}`
+      const allTimeCost = parseFloat((await client.get(totalCostKey)) || '0')
 
-      for (const key of allModelKeys) {
-        const modelMatch = key.match(/usage:.+:model:monthly:(.+):(\d{4}-\d{2})$/)
-        if (!modelMatch) {
-          continue
-        }
+      if (allTimeCost > 0) {
+        totalCost = allTimeCost
+        formattedCost = CostCalculator.formatCost(allTimeCost)
+        logger.debug(`📊 使用 allTimeCost 计算用户统计: ${allTimeCost}`)
+      } else {
+        // Fallback: 如果 allTimeCost 为空（旧键），尝试月度键
+        const allModelKeys = await client.keys(`usage:${keyId}:model:monthly:*:*`)
+        const modelUsageMap = new Map()
 
-        const model = modelMatch[1]
-        const data = await client.hgetall(key)
-
-        if (data && Object.keys(data).length > 0) {
-          if (!modelUsageMap.has(model)) {
-            modelUsageMap.set(model, {
-              inputTokens: 0,
-              outputTokens: 0,
-              cacheCreateTokens: 0,
-              cacheReadTokens: 0
-            })
+        for (const key of allModelKeys) {
+          const modelMatch = key.match(/usage:.+:model:monthly:(.+):(\d{4}-\d{2})$/)
+          if (!modelMatch) {
+            continue
           }
 
-          const modelUsage = modelUsageMap.get(model)
-          modelUsage.inputTokens += parseInt(data.inputTokens) || 0
-          modelUsage.outputTokens += parseInt(data.outputTokens) || 0
-          modelUsage.cacheCreateTokens += parseInt(data.cacheCreateTokens) || 0
-          modelUsage.cacheReadTokens += parseInt(data.cacheReadTokens) || 0
-        }
-      }
+          const model = modelMatch[1]
+          const data = await client.hgetall(key)
 
-      // 按模型计算费用并汇总
-      for (const [model, usage] of modelUsageMap) {
-        const usageData = {
-          input_tokens: usage.inputTokens,
-          output_tokens: usage.outputTokens,
-          cache_creation_input_tokens: usage.cacheCreateTokens,
-          cache_read_input_tokens: usage.cacheReadTokens
+          if (data && Object.keys(data).length > 0) {
+            if (!modelUsageMap.has(model)) {
+              modelUsageMap.set(model, {
+                inputTokens: 0,
+                outputTokens: 0,
+                cacheCreateTokens: 0,
+                cacheReadTokens: 0
+              })
+            }
+
+            const modelUsage = modelUsageMap.get(model)
+            modelUsage.inputTokens += parseInt(data.inputTokens) || 0
+            modelUsage.outputTokens += parseInt(data.outputTokens) || 0
+            modelUsage.cacheCreateTokens += parseInt(data.cacheCreateTokens) || 0
+            modelUsage.cacheReadTokens += parseInt(data.cacheReadTokens) || 0
+          }
         }
 
-        const costResult = CostCalculator.calculateCost(usageData, model)
-        totalCost += costResult.costs.total
-      }
+        // 按模型计算费用并汇总
+        for (const [model, usage] of modelUsageMap) {
+          const usageData = {
+            input_tokens: usage.inputTokens,
+            output_tokens: usage.outputTokens,
+            cache_creation_input_tokens: usage.cacheCreateTokens,
+            cache_read_input_tokens: usage.cacheReadTokens
+          }
 
-      // 如果没有模型级别的详细数据，回退到总体数据计算
-      if (modelUsageMap.size === 0 && fullKeyData.usage?.total?.allTokens > 0) {
-        const usage = fullKeyData.usage.total
-        const costUsage = {
-          input_tokens: usage.inputTokens || 0,
-          output_tokens: usage.outputTokens || 0,
-          cache_creation_input_tokens: usage.cacheCreateTokens || 0,
-          cache_read_input_tokens: usage.cacheReadTokens || 0
+          const costResult = CostCalculator.calculateCost(usageData, model)
+          totalCost += costResult.costs.total
         }
 
-        const costResult = CostCalculator.calculateCost(costUsage, 'claude-3-5-sonnet-20241022')
-        totalCost = costResult.costs.total
-      }
+        // 如果没有模型级别的详细数据，回退到总体数据计算
+        if (modelUsageMap.size === 0 && fullKeyData.usage?.total?.allTokens > 0) {
+          const usage = fullKeyData.usage.total
+          const costUsage = {
+            input_tokens: usage.inputTokens || 0,
+            output_tokens: usage.outputTokens || 0,
+            cache_creation_input_tokens: usage.cacheCreateTokens || 0,
+            cache_read_input_tokens: usage.cacheReadTokens || 0
+          }
 
-      formattedCost = CostCalculator.formatCost(totalCost)
+          const costResult = CostCalculator.calculateCost(costUsage, 'claude-3-5-sonnet-20241022')
+          totalCost = costResult.costs.total
+        }
+
+        formattedCost = CostCalculator.formatCost(totalCost)
+      }
     } catch (error) {
-      logger.warn(`Failed to calculate detailed cost for key ${keyId}:`, error)
+      logger.warn(`Failed to calculate cost for key ${keyId}:`, error)
       // 回退到简单计算
       if (fullKeyData.usage?.total?.allTokens > 0) {
         const usage = fullKeyData.usage.total
@@ -792,8 +804,8 @@ router.post('/api/batch-model-stats', async (req, res) => {
 
 // 🧪 API Key 端点测试接口 - 测试API Key是否能正常访问服务
 router.post('/api-key/test', async (req, res) => {
-  const axios = require('axios')
   const config = require('../../config/config')
+  const { sendStreamTestRequest } = require('../utils/testPayloadHelper')
 
   try {
     const { apiKey, model = 'claude-sonnet-4-5-20250929' } = req.body
@@ -805,7 +817,6 @@ router.post('/api-key/test', async (req, res) => {
       })
     }
 
-    // 基本格式验证
     if (typeof apiKey !== 'string' || apiKey.length < 10 || apiKey.length > 512) {
       return res.status(400).json({
         error: 'Invalid API key format',
@@ -813,7 +824,6 @@ router.post('/api-key/test', async (req, res) => {
       })
     }
 
-    // 首先验证API Key是否有效（不触发激活）
     const validation = await apiKeyService.validateApiKeyForStats(apiKey)
     if (!validation.valid) {
       return res.status(401).json({
@@ -824,244 +834,29 @@ router.post('/api-key/test', async (req, res) => {
 
     logger.api(`🧪 API Key test started for: ${validation.keyData.name} (${validation.keyData.id})`)
 
-    // 设置SSE响应头
-    res.setHeader('Content-Type', 'text/event-stream')
-    res.setHeader('Cache-Control', 'no-cache')
-    res.setHeader('Connection', 'keep-alive')
-    res.setHeader('X-Accel-Buffering', 'no')
-
-    // 发送测试开始事件
-    res.write(`data: ${JSON.stringify({ type: 'test_start', message: 'Test started' })}\n\n`)
-
-    // 构建测试请求，模拟 Claude CLI 客户端
     const port = config.server.port || 3000
-    const baseURL = `http://127.0.0.1:${port}`
+    const apiUrl = `http://127.0.0.1:${port}/api/v1/messages?beta=true`
 
-    const testPayload = {
-      model,
-      messages: [
-        {
-          role: 'user',
-          content: 'hi'
-        }
-      ],
-      system: [
-        {
-          type: 'text',
-          text: "You are Claude Code, Anthropic's official CLI for Claude."
-        }
-      ],
-      max_tokens: 32000,
-      temperature: 1,
-      stream: true
-    }
-
-    const headers = {
-      'Content-Type': 'application/json',
-      'User-Agent': 'claude-cli/2.0.52 (external, cli)',
-      'x-api-key': apiKey,
-      'anthropic-version': config.claude.apiVersion || '2023-06-01'
-    }
-
-    // 向自身服务发起测试请求
-    // 使用 validateStatus 允许所有状态码通过，以便我们可以处理流式错误响应
-    const response = await axios.post(`${baseURL}/api/v1/messages`, testPayload, {
-      headers,
-      responseType: 'stream',
-      timeout: 60000, // 60秒超时
-      validateStatus: () => true // 接受所有状态码，自行处理错误
-    })
-
-    // 检查响应状态码，如果不是2xx，尝试读取错误信息
-    if (response.status >= 400) {
-      logger.error(
-        `🧪 API Key test received error status ${response.status} for: ${validation.keyData.name}`
-      )
-
-      // 尝试从流中读取错误信息
-      let errorBody = ''
-      for await (const chunk of response.data) {
-        errorBody += chunk.toString()
-      }
-
-      let errorMessage = `HTTP ${response.status}`
-      try {
-        // 尝试解析SSE格式的错误
-        const lines = errorBody.split('\n')
-        for (const line of lines) {
-          if (line.startsWith('data: ')) {
-            const dataStr = line.substring(6).trim()
-            if (dataStr && dataStr !== '[DONE]') {
-              const data = JSON.parse(dataStr)
-              if (data.error?.message) {
-                errorMessage = data.error.message
-                break
-              } else if (data.message) {
-                errorMessage = data.message
-                break
-              } else if (typeof data.error === 'string') {
-                errorMessage = data.error
-                break
-              }
-            }
-          }
-        }
-        // 如果不是SSE格式，尝试直接解析JSON
-        if (errorMessage === `HTTP ${response.status}`) {
-          const jsonError = JSON.parse(errorBody)
-          errorMessage =
-            jsonError.error?.message || jsonError.message || jsonError.error || errorMessage
-        }
-      } catch {
-        // 解析失败，使用原始错误体或默认消息
-        if (errorBody && errorBody.length < 500) {
-          errorMessage = errorBody
-        }
-      }
-
-      res.write(`data: ${JSON.stringify({ type: 'error', error: errorMessage })}\n\n`)
-      res.write(
-        `data: ${JSON.stringify({ type: 'test_complete', success: false, error: errorMessage })}\n\n`
-      )
-      res.end()
-      return
-    }
-
-    let receivedContent = ''
-    let testSuccess = false
-    let upstreamError = null
-
-    // 处理流式响应
-    response.data.on('data', (chunk) => {
-      const lines = chunk.toString().split('\n')
-
-      for (const line of lines) {
-        if (line.startsWith('data: ')) {
-          const dataStr = line.substring(6).trim()
-          if (dataStr === '[DONE]') {
-            continue
-          }
-
-          try {
-            const data = JSON.parse(dataStr)
-
-            // 检查上游返回的错误事件
-            if (data.type === 'error' || data.error) {
-              let errorMsg = 'Unknown upstream error'
-
-              // 优先从 data.error 提取（如果是对象，获取其 message）
-              if (typeof data.error === 'object' && data.error?.message) {
-                errorMsg = data.error.message
-              } else if (typeof data.error === 'string' && data.error !== 'Claude API error') {
-                // 如果 error 是字符串且不是通用错误，直接使用
-                errorMsg = data.error
-              } else if (data.details) {
-                // 尝试从 details 字段解析详细错误（claudeRelayService 格式）
-                try {
-                  const details =
-                    typeof data.details === 'string' ? JSON.parse(data.details) : data.details
-                  if (details.error?.message) {
-                    errorMsg = details.error.message
-                  } else if (details.message) {
-                    errorMsg = details.message
-                  }
-                } catch {
-                  // details 不是有效 JSON，尝试直接使用
-                  if (typeof data.details === 'string' && data.details.length < 500) {
-                    errorMsg = data.details
-                  }
-                }
-              } else if (data.message) {
-                errorMsg = data.message
-              }
-
-              // 添加状态码信息（如果有）
-              if (data.status && errorMsg !== 'Unknown upstream error') {
-                errorMsg = `[${data.status}] ${errorMsg}`
-              }
-
-              upstreamError = errorMsg
-              logger.error(`🧪 Upstream error in test for: ${validation.keyData.name}:`, errorMsg)
-              res.write(`data: ${JSON.stringify({ type: 'error', error: errorMsg })}\n\n`)
-              continue
-            }
-
-            // 提取文本内容
-            if (data.type === 'content_block_delta' && data.delta?.text) {
-              receivedContent += data.delta.text
-              res.write(`data: ${JSON.stringify({ type: 'content', text: data.delta.text })}\n\n`)
-            }
-
-            // 消息结束
-            if (data.type === 'message_stop') {
-              testSuccess = true
-              res.write(`data: ${JSON.stringify({ type: 'message_stop' })}\n\n`)
-            }
-          } catch {
-            // 忽略解析错误
-          }
-        }
-      }
-    })
-
-    response.data.on('end', () => {
-      // 如果有上游错误，标记为失败
-      if (upstreamError) {
-        testSuccess = false
-      }
-
-      logger.api(
-        `🧪 API Key test completed for: ${validation.keyData.name}, success: ${testSuccess}, content length: ${receivedContent.length}${upstreamError ? `, error: ${upstreamError}` : ''}`
-      )
-      res.write(
-        `data: ${JSON.stringify({
-          type: 'test_complete',
-          success: testSuccess,
-          contentLength: receivedContent.length,
-          error: upstreamError || undefined
-        })}\n\n`
-      )
-      res.end()
-    })
-
-    response.data.on('error', (err) => {
-      logger.error(`🧪 API Key test stream error for: ${validation.keyData.name}`, err)
-
-      // 如果已经捕获了上游错误，优先使用那个
-      let errorMsg = upstreamError || err.message || 'Stream error'
-
-      // 如果错误消息是通用的 "Claude API error: xxx"，提供更友好的提示
-      if (errorMsg.startsWith('Claude API error:') && upstreamError) {
-        errorMsg = upstreamError
-      }
-
-      res.write(`data: ${JSON.stringify({ type: 'error', error: errorMsg })}\n\n`)
-      res.write(
-        `data: ${JSON.stringify({ type: 'test_complete', success: false, error: errorMsg })}\n\n`
-      )
-      res.end()
-    })
-
-    // 处理客户端断开连接
-    req.on('close', () => {
-      if (!res.writableEnded) {
-        response.data.destroy()
-      }
+    await sendStreamTestRequest({
+      apiUrl,
+      authorization: apiKey,
+      responseStream: res,
+      payload: createClaudeTestPayload(model, { stream: true }),
+      timeout: 60000,
+      extraHeaders: { 'x-api-key': apiKey }
     })
   } catch (error) {
     logger.error('❌ API Key test failed:', error)
 
-    // 如果还未发送响应头，返回JSON错误
     if (!res.headersSent) {
       return res.status(500).json({
         error: 'Test failed',
-        message: error.response?.data?.error?.message || error.message || 'Internal server error'
+        message: error.message || 'Internal server error'
       })
     }
 
-    // 如果已经是SSE流，发送错误事件
     res.write(
-      `data: ${JSON.stringify({ type: 'error', error: error.response?.data?.error?.message || error.message || 'Test failed' })}\n\n`
+      `data: ${JSON.stringify({ type: 'error', error: error.message || 'Test failed' })}\n\n`
     )
     res.end()
   }

src/routes/droidRoutes.js

@@ -4,12 +4,12 @@ const { authenticateApiKey } = require('../middleware/auth')
 const droidRelayService = require('../services/droidRelayService')
 const sessionHelper = require('../utils/sessionHelper')
 const logger = require('../utils/logger')
+const apiKeyService = require('../services/apiKeyService')
 
 const router = express.Router()
 
 function hasDroidPermission(apiKeyData) {
-  const permissions = apiKeyData?.permissions || 'all'
-  return permissions === 'all' || permissions === 'droid'
+  return apiKeyService.hasPermission(apiKeyData?.permissions, 'droid')
 }
 
 /**

src/routes/geminiRoutes.js

@@ -29,6 +29,7 @@ const {
   handleStreamGenerateContent,
   handleLoadCodeAssist,
   handleOnboardUser,
+  handleRetrieveUserQuota,
   handleCountTokens,
   handleStandardGenerateContent,
   handleStandardStreamGenerateContent,
@@ -68,7 +69,7 @@ router.get('/usage', authenticateApiKey, handleUsage)
 router.get('/key-info', authenticateApiKey, handleKeyInfo)
 
 // ============================================================================
-// v1internal 独有路由（listExperiments）
+// v1internal 独有路由
 // ============================================================================
 
 /**
@@ -81,6 +82,12 @@ router.post(
   handleSimpleEndpoint('listExperiments')
 )
 
+/**
+ * POST /v1internal:retrieveUserQuota
+ * 获取用户配额信息（Gemini CLI 0.22.2+ 需要）
+ */
+router.post('/v1internal\\:retrieveUserQuota', authenticateApiKey, handleRetrieveUserQuota)
+
 /**
  * POST /v1beta/models/:modelName:listExperiments
  * 带模型参数的实验列表（只有 geminiRoutes 定义此路由）

src/routes/openaiClaudeRoutes.js

@@ -8,6 +8,7 @@ const router = express.Router()
 const logger = require('../utils/logger')
 const { authenticateApiKey } = require('../middleware/auth')
 const claudeRelayService = require('../services/claudeRelayService')
+const claudeConsoleRelayService = require('../services/claudeConsoleRelayService')
 const openaiToClaude = require('../services/openaiToClaude')
 const apiKeyService = require('../services/apiKeyService')
 const unifiedClaudeScheduler = require('../services/unifiedClaudeScheduler')
@@ -15,11 +16,11 @@ const claudeCodeHeadersService = require('../services/claudeCodeHeadersService')
 const sessionHelper = require('../utils/sessionHelper')
 const { updateRateLimitCounters } = require('../utils/rateLimitHelper')
 const pricingService = require('../services/pricingService')
+const { getEffectiveModel } = require('../utils/modelHelper')
 
 // 🔧 辅助函数：检查 API Key 权限
 function checkPermissions(apiKeyData, requiredPermission = 'claude') {
-  const permissions = apiKeyData.permissions || 'all'
-  return permissions === 'all' || permissions === requiredPermission
+  return apiKeyService.hasPermission(apiKeyData?.permissions, requiredPermission)
 }
 
 function queueRateLimitUpdate(rateLimitInfo, usageSummary, model, context = '') {
@@ -75,9 +76,9 @@ router.get('/v1/models', authenticateApiKey, async (req, res) => {
       }
     ]
 
-    // 如果启用了模型限制，过滤模型列表
+    // 如果启用了模型限制，视为黑名单：过滤掉受限模型
     if (apiKeyData.enableModelRestriction && apiKeyData.restrictedModels?.length > 0) {
-      models = models.filter((model) => apiKeyData.restrictedModels.includes(model.id))
+      models = models.filter((model) => !apiKeyData.restrictedModels.includes(model.id))
     }
 
     res.json({
@@ -114,9 +115,9 @@ router.get('/v1/models/:model', authenticateApiKey, async (req, res) => {
       })
     }
 
-    // 检查模型限制
+    // 模型限制（黑名单）：命中则直接拒绝
     if (apiKeyData.enableModelRestriction && apiKeyData.restrictedModels?.length > 0) {
-      if (!apiKeyData.restrictedModels.includes(modelId)) {
+      if (apiKeyData.restrictedModels.includes(modelId)) {
         return res.status(404).json({
           error: {
             message: `Model '${modelId}' not found`,
@@ -199,9 +200,10 @@ async function handleChatCompletion(req, res, apiKeyData) {
     // 转换 OpenAI 请求为 Claude 格式
     const claudeRequest = openaiToClaude.convertRequest(req.body)
 
-    // 检查模型限制
+    // 模型限制（黑名单）：命中受限模型则拒绝
     if (apiKeyData.enableModelRestriction && apiKeyData.restrictedModels?.length > 0) {
-      if (!apiKeyData.restrictedModels.includes(claudeRequest.model)) {
+      const effectiveModel = getEffectiveModel(claudeRequest.model || '')
+      if (apiKeyData.restrictedModels.includes(effectiveModel)) {
         return res.status(403).json({
           error: {
             message: `Model ${req.body.model} is not allowed for this API key`,
@@ -233,7 +235,7 @@ async function handleChatCompletion(req, res, apiKeyData) {
       }
       throw error
     }
-    const { accountId } = accountSelection
+    const { accountId, accountType } = accountSelection
 
     // 获取该账号存储的 Claude Code headers
     const claudeCodeHeaders = await claudeCodeHeadersService.getAccountHeaders(accountId)
@@ -263,72 +265,105 @@ async function handleChatCompletion(req, res, apiKeyData) {
         }
       })
 
-      // 使用转换后的响应流 (使用 OAuth-only beta header，添加 Claude Code 必需的 headers)
-      await claudeRelayService.relayStreamRequestWithUsageCapture(
-        claudeRequest,
-        apiKeyData,
-        res,
-        claudeCodeHeaders,
-        (usage) => {
-          // 记录使用统计
-          if (usage && usage.input_tokens !== undefined && usage.output_tokens !== undefined) {
-            const model = usage.model || claudeRequest.model
-            const cacheCreateTokens =
-              (usage.cache_creation && typeof usage.cache_creation === 'object'
-                ? (usage.cache_creation.ephemeral_5m_input_tokens || 0) +
-                  (usage.cache_creation.ephemeral_1h_input_tokens || 0)
-                : usage.cache_creation_input_tokens || 0) || 0
-            const cacheReadTokens = usage.cache_read_input_tokens || 0
+      // 使用转换后的响应流 (根据账户类型选择转发服务)
+      // 创建 usage 回调函数
+      const usageCallback = (usage) => {
+        // 记录使用统计
+        if (usage && usage.input_tokens !== undefined && usage.output_tokens !== undefined) {
+          const model = usage.model || claudeRequest.model
+          const cacheCreateTokens =
+            (usage.cache_creation && typeof usage.cache_creation === 'object'
+              ? (usage.cache_creation.ephemeral_5m_input_tokens || 0) +
+                (usage.cache_creation.ephemeral_1h_input_tokens || 0)
+              : usage.cache_creation_input_tokens || 0) || 0
+          const cacheReadTokens = usage.cache_read_input_tokens || 0
 
-            // 使用新的 recordUsageWithDetails 方法来支持详细的缓存数据
-            apiKeyService
-              .recordUsageWithDetails(
-                apiKeyData.id,
-                usage, // 直接传递整个 usage 对象，包含可能的 cache_creation 详细数据
-                model,
-                accountId
-              )
-              .catch((error) => {
-                logger.error('❌ Failed to record usage:', error)
-              })
-
-            queueRateLimitUpdate(
-              req.rateLimitInfo,
-              {
-                inputTokens: usage.input_tokens || 0,
-                outputTokens: usage.output_tokens || 0,
-                cacheCreateTokens,
-                cacheReadTokens
-              },
+          // 使用新的 recordUsageWithDetails 方法来支持详细的缓存数据
+          apiKeyService
+            .recordUsageWithDetails(
+              apiKeyData.id,
+              usage, // 直接传递整个 usage 对象，包含可能的 cache_creation 详细数据
               model,
-              'openai-claude-stream'
+              accountId,
+              accountType
             )
-          }
-        },
-        // 流转换器
-        (() => {
-          // 为每个请求创建独立的会话ID
-          const sessionId = `chatcmpl-${Math.random().toString(36).substring(2, 15)}${Math.random().toString(36).substring(2, 15)}`
-          return (chunk) => openaiToClaude.convertStreamChunk(chunk, req.body.model, sessionId)
-        })(),
-        {
-          betaHeader:
-            'oauth-2025-04-20,claude-code-20250219,interleaved-thinking-2025-05-14,fine-grained-tool-streaming-2025-05-14'
+            .catch((error) => {
+              logger.error('❌ Failed to record usage:', error)
+            })
+
+          queueRateLimitUpdate(
+            req.rateLimitInfo,
+            {
+              inputTokens: usage.input_tokens || 0,
+              outputTokens: usage.output_tokens || 0,
+              cacheCreateTokens,
+              cacheReadTokens
+            },
+            model,
+            `openai-${accountType}-stream`
+          )
         }
-      )
+      }
+
+      // 创建流转换器
+      const sessionId = `chatcmpl-${Math.random().toString(36).substring(2, 15)}${Math.random().toString(36).substring(2, 15)}`
+      const streamTransformer = (chunk) =>
+        openaiToClaude.convertStreamChunk(chunk, req.body.model, sessionId)
+
+      // 根据账户类型选择转发服务
+      if (accountType === 'claude-console') {
+        // Claude Console 账户使用 Console 转发服务
+        await claudeConsoleRelayService.relayStreamRequestWithUsageCapture(
+          claudeRequest,
+          apiKeyData,
+          res,
+          claudeCodeHeaders,
+          usageCallback,
+          accountId,
+          streamTransformer
+        )
+      } else {
+        // Claude Official 账户使用标准转发服务
+        await claudeRelayService.relayStreamRequestWithUsageCapture(
+          claudeRequest,
+          apiKeyData,
+          res,
+          claudeCodeHeaders,
+          usageCallback,
+          streamTransformer,
+          {
+            betaHeader:
+              'oauth-2025-04-20,claude-code-20250219,interleaved-thinking-2025-05-14,fine-grained-tool-streaming-2025-05-14'
+          }
+        )
+      }
     } else {
       // 非流式请求
       logger.info(`📄 Processing OpenAI non-stream request for model: ${req.body.model}`)
 
-      // 发送请求到 Claude (使用 OAuth-only beta header，添加 Claude Code 必需的 headers)
-      const claudeResponse = await claudeRelayService.relayRequest(
-        claudeRequest,
-        apiKeyData,
-        req,
-        res,
-        claudeCodeHeaders,
-        { betaHeader: 'oauth-2025-04-20' }
-      )
+      // 根据账户类型选择转发服务
+      let claudeResponse
+      if (accountType === 'claude-console') {
+        // Claude Console 账户使用 Console 转发服务
+        claudeResponse = await claudeConsoleRelayService.relayRequest(
+          claudeRequest,
+          apiKeyData,
+          req,
+          res,
+          claudeCodeHeaders,
+          accountId
+        )
+      } else {
+        // Claude Official 账户使用标准转发服务
+        claudeResponse = await claudeRelayService.relayRequest(
+          claudeRequest,
+          apiKeyData,
+          req,
+          res,
+          claudeCodeHeaders,
+          { betaHeader: 'oauth-2025-04-20' }
+        )
+      }
 
       // 解析 Claude 响应
       let claudeData
@@ -374,7 +409,8 @@ async function handleChatCompletion(req, res, apiKeyData) {
             apiKeyData.id,
             usage, // 直接传递整个 usage 对象，包含可能的 cache_creation 详细数据
             claudeRequest.model,
-            accountId
+            accountId,
+            accountType
           )
           .catch((error) => {
             logger.error('❌ Failed to record usage:', error)
@@ -389,7 +425,7 @@ async function handleChatCompletion(req, res, apiKeyData) {
             cacheReadTokens
           },
           claudeRequest.model,
-          'openai-claude-non-stream'
+          `openai-${accountType}-non-stream`
         )
       }
 
@@ -400,16 +436,29 @@ async function handleChatCompletion(req, res, apiKeyData) {
     const duration = Date.now() - startTime
     logger.info(`✅ OpenAI-Claude request completed in ${duration}ms`)
   } catch (error) {
-    logger.error('❌ OpenAI-Claude request error:', error)
+    // 客户端主动断开连接是正常情况，使用 INFO 级别
+    if (error.message === 'Client disconnected') {
+      logger.info('🔌 OpenAI-Claude stream ended: Client disconnected')
+    } else {
+      logger.error('❌ OpenAI-Claude request error:', error)
+    }
 
-    const status = error.status || 500
-    res.status(status).json({
-      error: {
-        message: error.message || 'Internal server error',
-        type: 'server_error',
-        code: 'internal_error'
+    // 检查响应是否已发送（流式响应场景），避免 ERR_HTTP_HEADERS_SENT
+    if (!res.headersSent) {
+      // 客户端断开使用 499 状态码 (Client Closed Request)
+      if (error.message === 'Client disconnected') {
+        res.status(499).end()
+      } else {
+        const status = error.status || 500
+        res.status(status).json({
+          error: {
+            message: error.message || 'Internal server error',
+            type: 'server_error',
+            code: 'internal_error'
+          }
+        })
       }
-    })
+    }
   } finally {
     // 清理资源
     if (abortController) {

src/routes/openaiGeminiRoutes.js

@@ -6,6 +6,7 @@ const geminiAccountService = require('../services/geminiAccountService')
 const unifiedGeminiScheduler = require('../services/unifiedGeminiScheduler')
 const { getAvailableModels } = require('../services/geminiRelayService')
 const crypto = require('crypto')
+const apiKeyService = require('../services/apiKeyService')
 
 // 生成会话哈希
 function generateSessionHash(req) {
@@ -19,10 +20,19 @@ function generateSessionHash(req) {
   return crypto.createHash('sha256').update(sessionData).digest('hex')
 }
 
+function ensureAntigravityProjectId(account) {
+  if (account.projectId) {
+    return account.projectId
+  }
+  if (account.tempProjectId) {
+    return account.tempProjectId
+  }
+  return `ag-${crypto.randomBytes(8).toString('hex')}`
+}
+
 // 检查 API Key 权限
 function checkPermissions(apiKeyData, requiredPermission = 'gemini') {
-  const permissions = apiKeyData.permissions || 'all'
-  return permissions === 'all' || permissions === requiredPermission
+  return apiKeyService.hasPermission(apiKeyData?.permissions, requiredPermission)
 }
 
 // 转换 OpenAI 消息格式到 Gemini 格式
@@ -335,25 +345,48 @@ router.post('/v1/chat/completions', authenticateApiKey, async (req, res) => {
     const client = await geminiAccountService.getOauthClient(
       account.accessToken,
       account.refreshToken,
-      proxyConfig
+      proxyConfig,
+      account.oauthProvider
     )
     if (actualStream) {
       // 流式响应
+      const oauthProvider = account.oauthProvider || 'gemini-cli'
+      let { projectId } = account
+
+      if (oauthProvider === 'antigravity') {
+        projectId = ensureAntigravityProjectId(account)
+        if (!account.projectId && account.tempProjectId !== projectId) {
+          await geminiAccountService.updateTempProjectId(account.id, projectId)
+          account.tempProjectId = projectId
+        }
+      }
+
       logger.info('StreamGenerateContent request', {
         model,
-        projectId: account.projectId,
+        projectId,
         apiKeyId: apiKeyData.id
       })
 
-      const streamResponse = await geminiAccountService.generateContentStream(
-        client,
-        { model, request: geminiRequestBody },
-        null, // user_prompt_id
-        account.projectId, // 使用有权限的项目ID
-        apiKeyData.id, // 使用 API Key ID 作为 session ID
-        abortController.signal, // 传递中止信号
-        proxyConfig // 传递代理配置
-      )
+      const streamResponse =
+        oauthProvider === 'antigravity'
+          ? await geminiAccountService.generateContentStreamAntigravity(
+              client,
+              { model, request: geminiRequestBody },
+              null, // user_prompt_id
+              projectId,
+              apiKeyData.id, // 使用 API Key ID 作为 session ID
+              abortController.signal, // 传递中止信号
+              proxyConfig // 传递代理配置
+            )
+          : await geminiAccountService.generateContentStream(
+              client,
+              { model, request: geminiRequestBody },
+              null, // user_prompt_id
+              projectId, // 使用有权限的项目ID
+              apiKeyData.id, // 使用 API Key ID 作为 session ID
+              abortController.signal, // 传递中止信号
+              proxyConfig // 传递代理配置
+            )
 
       // 设置流式响应头
       res.setHeader('Content-Type', 'text/event-stream')
@@ -499,7 +532,6 @@ router.post('/v1/chat/completions', authenticateApiKey, async (req, res) => {
         // 记录使用统计
         if (!usageReported && totalUsage.totalTokenCount > 0) {
           try {
-            const apiKeyService = require('../services/apiKeyService')
             await apiKeyService.recordUsage(
               apiKeyData.id,
               totalUsage.promptTokenCount || 0,
@@ -559,20 +591,41 @@ router.post('/v1/chat/completions', authenticateApiKey, async (req, res) => {
       })
     } else {
       // 非流式响应
+      const oauthProvider = account.oauthProvider || 'gemini-cli'
+      let { projectId } = account
+
+      if (oauthProvider === 'antigravity') {
+        projectId = ensureAntigravityProjectId(account)
+        if (!account.projectId && account.tempProjectId !== projectId) {
+          await geminiAccountService.updateTempProjectId(account.id, projectId)
+          account.tempProjectId = projectId
+        }
+      }
+
       logger.info('GenerateContent request', {
         model,
-        projectId: account.projectId,
+        projectId,
         apiKeyId: apiKeyData.id
       })
 
-      const response = await geminiAccountService.generateContent(
-        client,
-        { model, request: geminiRequestBody },
-        null, // user_prompt_id
-        account.projectId, // 使用有权限的项目ID
-        apiKeyData.id, // 使用 API Key ID 作为 session ID
-        proxyConfig // 传递代理配置
-      )
+      const response =
+        oauthProvider === 'antigravity'
+          ? await geminiAccountService.generateContentAntigravity(
+              client,
+              { model, request: geminiRequestBody },
+              null, // user_prompt_id
+              projectId,
+              apiKeyData.id, // 使用 API Key ID 作为 session ID
+              proxyConfig // 传递代理配置
+            )
+          : await geminiAccountService.generateContent(
+              client,
+              { model, request: geminiRequestBody },
+              null, // user_prompt_id
+              projectId, // 使用有权限的项目ID
+              apiKeyData.id, // 使用 API Key ID 作为 session ID
+              proxyConfig // 传递代理配置
+            )
 
       // 转换为 OpenAI 格式并返回
       const openaiResponse = convertGeminiResponseToOpenAI(response, model, false)
@@ -580,7 +633,6 @@ router.post('/v1/chat/completions', authenticateApiKey, async (req, res) => {
       // 记录使用统计
       if (openaiResponse.usage) {
         try {
-          const apiKeyService = require('../services/apiKeyService')
           await apiKeyService.recordUsage(
             apiKeyData.id,
             openaiResponse.usage.prompt_tokens || 0,
@@ -604,7 +656,15 @@ router.post('/v1/chat/completions', authenticateApiKey, async (req, res) => {
     const duration = Date.now() - startTime
     logger.info(`OpenAI-Gemini request completed in ${duration}ms`)
   } catch (error) {
-    logger.error('OpenAI-Gemini request error:', error)
+    const statusForLog = error?.status || error?.response?.status
+    logger.error('OpenAI-Gemini request error', {
+      message: error?.message,
+      status: statusForLog,
+      code: error?.code,
+      requestUrl: error?.config?.url,
+      requestMethod: error?.config?.method,
+      upstreamTraceId: error?.response?.headers?.['x-cloudaicompanion-trace-id']
+    })
 
     // 处理速率限制
     if (error.status === 429) {
@@ -613,17 +673,24 @@ router.post('/v1/chat/completions', authenticateApiKey, async (req, res) => {
       }
     }
 
-    // 返回 OpenAI 格式的错误响应
-    const status = error.status || 500
-    const errorResponse = {
-      error: error.error || {
-        message: error.message || 'Internal server error',
-        type: 'server_error',
-        code: 'internal_error'
+    // 检查响应是否已发送（流式响应场景），避免 ERR_HTTP_HEADERS_SENT
+    if (!res.headersSent) {
+      // 客户端断开使用 499 状态码 (Client Closed Request)
+      if (error.message === 'Client disconnected') {
+        res.status(499).end()
+      } else {
+        // 返回 OpenAI 格式的错误响应
+        const status = error.status || 500
+        const errorResponse = {
+          error: error.error || {
+            message: error.message || 'Internal server error',
+            type: 'server_error',
+            code: 'internal_error'
+          }
+        }
+        res.status(status).json(errorResponse)
       }
     }
-
-    res.status(status).json(errorResponse)
   } finally {
     // 清理资源
     if (abortController) {
@@ -633,8 +700,8 @@ router.post('/v1/chat/completions', authenticateApiKey, async (req, res) => {
   return undefined
 })
 
-// OpenAI 兼容的模型列表端点
-router.get('/v1/models', authenticateApiKey, async (req, res) => {
+// 获取可用模型列表的共享处理器
+async function handleGetModels(req, res) {
   try {
     const apiKeyData = req.apiKey
 
@@ -665,8 +732,21 @@ router.get('/v1/models', authenticateApiKey, async (req, res) => {
     let models = []
 
     if (account) {
-      // 获取实际的模型列表
-      models = await getAvailableModels(account.accessToken, account.proxy)
+      // 获取实际的模型列表（失败时回退到默认列表，避免影响 /v1/models 可用性）
+      try {
+        const oauthProvider = account.oauthProvider || 'gemini-cli'
+        models =
+          oauthProvider === 'antigravity'
+            ? await geminiAccountService.fetchAvailableModelsAntigravity(
+                account.accessToken,
+                account.proxy,
+                account.refreshToken
+              )
+            : await getAvailableModels(account.accessToken, account.proxy)
+      } catch (error) {
+        logger.warn('Failed to get Gemini models list from upstream, fallback to default:', error)
+        models = []
+      }
     } else {
       // 返回默认模型列表
       models = [
@@ -679,6 +759,17 @@ router.get('/v1/models', authenticateApiKey, async (req, res) => {
       ]
     }
 
+    if (!models || models.length === 0) {
+      models = [
+        {
+          id: 'gemini-2.0-flash-exp',
+          object: 'model',
+          created: Math.floor(Date.now() / 1000),
+          owned_by: 'google'
+        }
+      ]
+    }
+
     // 如果启用了模型限制，过滤模型列表
     if (apiKeyData.enableModelRestriction && apiKeyData.restrictedModels.length > 0) {
       models = models.filter((model) => apiKeyData.restrictedModels.includes(model.id))
@@ -698,8 +789,13 @@ router.get('/v1/models', authenticateApiKey, async (req, res) => {
       }
     })
   }
-  return undefined
-})
+}
+
+// OpenAI 兼容的模型列表端点 (带 v1 版)
+router.get('/v1/models', authenticateApiKey, handleGetModels)
+
+// OpenAI 兼容的模型列表端点 (根路径版，方便第三方加载)
+router.get('/models', authenticateApiKey, handleGetModels)
 
 // OpenAI 兼容的模型详情端点
 router.get('/v1/models/:model', authenticateApiKey, async (req, res) => {

src/routes/openaiRoutes.js

@@ -20,8 +20,7 @@ function createProxyAgent(proxy) {
 
 // 检查 API Key 是否具备 OpenAI 权限
 function checkOpenAIPermissions(apiKeyData) {
-  const permissions = apiKeyData?.permissions || 'all'
-  return permissions === 'all' || permissions === 'openai'
+  return apiKeyService.hasPermission(apiKeyData?.permissions, 'openai')
 }
 
 function normalizeHeaders(headers = {}) {
@@ -247,9 +246,11 @@ const handleResponses = async (req, res) => {
 
     // 从请求体中提取模型和流式标志
     let requestedModel = req.body?.model || null
+    const isCodexModel =
+      typeof requestedModel === 'string' && requestedModel.toLowerCase().includes('codex')
 
-    // 如果模型是 gpt-5 开头且后面还有内容（如 gpt-5-2025-08-07），则覆盖为 gpt-5
-    if (requestedModel && requestedModel.startsWith('gpt-5-') && requestedModel !== 'gpt-5-codex') {
+    // 如果模型是 gpt-5 开头且后面还有内容（如 gpt-5-2025-08-07），并且不是 Codex 系列，则覆盖为 gpt-5
+    if (requestedModel && requestedModel.startsWith('gpt-5-') && !isCodexModel) {
       logger.info(`📝 Model ${requestedModel} detected, normalizing to gpt-5 for Codex API`)
       requestedModel = 'gpt-5'
       req.body.model = 'gpt-5' // 同时更新请求体中的模型
@@ -273,7 +274,9 @@ const handleResponses = async (req, res) => {
         'text_formatting',
         'truncation',
         'text',
-        'service_tier'
+        'service_tier',
+        'prompt_cache_retention',
+        'safety_identifier'
       ]
       fieldsToRemove.forEach((field) => {
         delete req.body[field]
@@ -903,7 +906,7 @@ router.get('/key-info', authenticateApiKey, async (req, res) => {
       id: keyData.id,
       name: keyData.name,
       description: keyData.description,
-      permissions: keyData.permissions || 'all',
+      permissions: keyData.permissions,
       token_limit: keyData.tokenLimit,
       tokens_used: keyData.usage.total.tokens,
       tokens_remaining:

src/routes/unified.js

@@ -8,6 +8,7 @@ const {
   handleStreamGenerateContent: geminiHandleStreamGenerateContent
 } = require('../handlers/geminiHandlers')
 const openaiRoutes = require('./openaiRoutes')
+const apiKeyService = require('../services/apiKeyService')
 
 const router = express.Router()
 
@@ -45,11 +46,11 @@ async function routeToBackend(req, res, requestedModel) {
   logger.info(`🔀 Routing request - Model: ${requestedModel}, Backend: ${backend}`)
 
   // 检查权限
-  const permissions = req.apiKey.permissions || 'all'
+  const { permissions } = req.apiKey
 
   if (backend === 'claude') {
     // Claude 后端：通过 OpenAI 兼容层
-    if (permissions !== 'all' && permissions !== 'claude') {
+    if (!apiKeyService.hasPermission(permissions, 'claude')) {
       return res.status(403).json({
         error: {
           message: 'This API key does not have permission to access Claude',
@@ -61,7 +62,7 @@ async function routeToBackend(req, res, requestedModel) {
     await handleChatCompletion(req, res, req.apiKey)
   } else if (backend === 'openai') {
     // OpenAI 后端
-    if (permissions !== 'all' && permissions !== 'openai') {
+    if (!apiKeyService.hasPermission(permissions, 'openai')) {
       return res.status(403).json({
         error: {
           message: 'This API key does not have permission to access OpenAI',
@@ -73,7 +74,7 @@ async function routeToBackend(req, res, requestedModel) {
     return await openaiRoutes.handleResponses(req, res)
   } else if (backend === 'gemini') {
     // Gemini 后端
-    if (permissions !== 'all' && permissions !== 'gemini') {
+    if (!apiKeyService.hasPermission(permissions, 'gemini')) {
       return res.status(403).json({
         error: {
           message: 'This API key does not have permission to access Gemini',

src/routes/web.js

@@ -164,13 +164,27 @@ router.post('/auth/change-password', async (req, res) => {
 
     // 获取当前会话
     const sessionData = await redis.getSession(token)
-    if (!sessionData) {
+
+    // 🔒 安全修复：检查空对象
+    if (!sessionData || Object.keys(sessionData).length === 0) {
       return res.status(401).json({
         error: 'Invalid token',
         message: 'Session expired or invalid'
       })
     }
 
+    // 🔒 安全修复：验证会话完整性
+    if (!sessionData.username || !sessionData.loginTime) {
+      logger.security(
+        `🔒 Invalid session structure in /auth/change-password from ${req.ip || 'unknown'}`
+      )
+      await redis.deleteSession(token)
+      return res.status(401).json({
+        error: 'Invalid session',
+        message: 'Session data corrupted or incomplete'
+      })
+    }
+
     // 获取当前管理员信息
     const adminData = await redis.getSession('admin_credentials')
     if (!adminData) {
@@ -269,13 +283,25 @@ router.get('/auth/user', async (req, res) => {
 
     // 获取当前会话
     const sessionData = await redis.getSession(token)
-    if (!sessionData) {
+
+    // 🔒 安全修复：检查空对象
+    if (!sessionData || Object.keys(sessionData).length === 0) {
       return res.status(401).json({
         error: 'Invalid token',
         message: 'Session expired or invalid'
       })
     }
 
+    // 🔒 安全修复：验证会话完整性
+    if (!sessionData.username || !sessionData.loginTime) {
+      logger.security(`🔒 Invalid session structure in /auth/user from ${req.ip || 'unknown'}`)
+      await redis.deleteSession(token)
+      return res.status(401).json({
+        error: 'Invalid session',
+        message: 'Session data corrupted or incomplete'
+      })
+    }
+
     // 获取管理员信息
     const adminData = await redis.getSession('admin_credentials')
     if (!adminData) {
@@ -316,13 +342,24 @@ router.post('/auth/refresh', async (req, res) => {
 
     const sessionData = await redis.getSession(token)
 
-    if (!sessionData) {
+    // 🔒 安全修复：检查空对象（hgetall 对不存在的 key 返回 {}）
+    if (!sessionData || Object.keys(sessionData).length === 0) {
       return res.status(401).json({
         error: 'Invalid token',
         message: 'Session expired or invalid'
       })
     }
 
+    // 🔒 安全修复：验证会话完整性（必须有 username 和 loginTime）
+    if (!sessionData.username || !sessionData.loginTime) {
+      logger.security(`🔒 Invalid session structure detected from ${req.ip || 'unknown'}`)
+      await redis.deleteSession(token) // 清理无效/伪造的会话
+      return res.status(401).json({
+        error: 'Invalid session',
+        message: 'Session data corrupted or incomplete'
+      })
+    }
+
     // 更新最后活动时间
     sessionData.lastActivity = new Date().toISOString()
     await redis.setSession(token, sessionData, config.security.adminSessionTimeout)

src/services/accountBalanceService.js

@@ -0,0 +1,789 @@
+const redis = require('../models/redis')
+const balanceScriptService = require('./balanceScriptService')
+const logger = require('../utils/logger')
+const CostCalculator = require('../utils/costCalculator')
+const { isBalanceScriptEnabled } = require('../utils/featureFlags')
+
+class AccountBalanceService {
+  constructor(options = {}) {
+    this.redis = options.redis || redis
+    this.logger = options.logger || logger
+
+    this.providers = new Map()
+
+    this.CACHE_TTL_SECONDS = 3600
+    this.LOCAL_TTL_SECONDS = 300
+
+    this.LOW_BALANCE_THRESHOLD = 10
+    this.HIGH_USAGE_THRESHOLD_PERCENT = 90
+    this.DEFAULT_CONCURRENCY = 10
+  }
+
+  getSupportedPlatforms() {
+    return [
+      'claude',
+      'claude-console',
+      'gemini',
+      'gemini-api',
+      'openai',
+      'openai-responses',
+      'azure_openai',
+      'bedrock',
+      'droid',
+      'ccr'
+    ]
+  }
+
+  normalizePlatform(platform) {
+    if (!platform) {
+      return null
+    }
+
+    const value = String(platform).trim().toLowerCase()
+
+    // 兼容实施文档与历史命名
+    if (value === 'claude-official') {
+      return 'claude'
+    }
+    if (value === 'azure-openai') {
+      return 'azure_openai'
+    }
+
+    // 保持前端平台键一致
+    return value
+  }
+
+  registerProvider(platform, provider) {
+    const normalized = this.normalizePlatform(platform)
+    if (!normalized) {
+      throw new Error('registerProvider: 缺少 platform')
+    }
+    if (!provider || typeof provider.queryBalance !== 'function') {
+      throw new Error(`registerProvider: Provider 无效 (${normalized})`)
+    }
+    this.providers.set(normalized, provider)
+  }
+
+  async getAccountBalance(accountId, platform, options = {}) {
+    const normalizedPlatform = this.normalizePlatform(platform)
+    const account = await this.getAccount(accountId, normalizedPlatform)
+    if (!account) {
+      return null
+    }
+    return await this._getAccountBalanceForAccount(account, normalizedPlatform, options)
+  }
+
+  async refreshAccountBalance(accountId, platform) {
+    const normalizedPlatform = this.normalizePlatform(platform)
+    const account = await this.getAccount(accountId, normalizedPlatform)
+    if (!account) {
+      return null
+    }
+
+    return await this._getAccountBalanceForAccount(account, normalizedPlatform, {
+      queryApi: true,
+      useCache: false
+    })
+  }
+
+  async getAllAccountsBalance(platform, options = {}) {
+    const normalizedPlatform = this.normalizePlatform(platform)
+    const accounts = await this.getAllAccountsByPlatform(normalizedPlatform)
+    const queryApi = this._parseBoolean(options.queryApi) || false
+    const useCache = options.useCache !== false
+
+    const results = await this._mapWithConcurrency(
+      accounts,
+      this.DEFAULT_CONCURRENCY,
+      async (acc) => {
+        try {
+          const balance = await this._getAccountBalanceForAccount(acc, normalizedPlatform, {
+            queryApi,
+            useCache
+          })
+          return { ...balance, name: acc.name || '' }
+        } catch (error) {
+          this.logger.error(`批量获取余额失败: ${normalizedPlatform}:${acc?.id}`, error)
+          return {
+            success: true,
+            data: {
+              accountId: acc?.id,
+              platform: normalizedPlatform,
+              balance: null,
+              quota: null,
+              statistics: {},
+              source: 'local',
+              lastRefreshAt: new Date().toISOString(),
+              cacheExpiresAt: null,
+              status: 'error',
+              error: error.message || '批量查询失败'
+            },
+            name: acc?.name || ''
+          }
+        }
+      }
+    )
+
+    return results
+  }
+
+  async getBalanceSummary() {
+    const platforms = this.getSupportedPlatforms()
+
+    const summary = {
+      totalBalance: 0,
+      totalCost: 0,
+      lowBalanceCount: 0,
+      platforms: {}
+    }
+
+    for (const platform of platforms) {
+      const accounts = await this.getAllAccountsByPlatform(platform)
+      const platformData = {
+        count: accounts.length,
+        totalBalance: 0,
+        totalCost: 0,
+        lowBalanceCount: 0,
+        accounts: []
+      }
+
+      const balances = await this._mapWithConcurrency(
+        accounts,
+        this.DEFAULT_CONCURRENCY,
+        async (acc) => {
+          const balance = await this._getAccountBalanceForAccount(acc, platform, {
+            queryApi: false,
+            useCache: true
+          })
+          return { ...balance, name: acc.name || '' }
+        }
+      )
+
+      for (const item of balances) {
+        platformData.accounts.push(item)
+
+        const amount = item?.data?.balance?.amount
+        const percentage = item?.data?.quota?.percentage
+        const totalCost = Number(item?.data?.statistics?.totalCost || 0)
+
+        const hasAmount = typeof amount === 'number' && Number.isFinite(amount)
+        const isLowBalance = hasAmount && amount < this.LOW_BALANCE_THRESHOLD
+        const isHighUsage =
+          typeof percentage === 'number' &&
+          Number.isFinite(percentage) &&
+          percentage > this.HIGH_USAGE_THRESHOLD_PERCENT
+
+        if (hasAmount) {
+          platformData.totalBalance += amount
+        }
+
+        if (isLowBalance || isHighUsage) {
+          platformData.lowBalanceCount += 1
+          summary.lowBalanceCount += 1
+        }
+
+        platformData.totalCost += totalCost
+      }
+
+      summary.platforms[platform] = platformData
+      summary.totalBalance += platformData.totalBalance
+      summary.totalCost += platformData.totalCost
+    }
+
+    return summary
+  }
+
+  async clearCache(accountId, platform) {
+    const normalizedPlatform = this.normalizePlatform(platform)
+    if (!normalizedPlatform) {
+      throw new Error('缺少 platform 参数')
+    }
+
+    await this.redis.deleteAccountBalance(normalizedPlatform, accountId)
+    this.logger.info(`余额缓存已清除: ${normalizedPlatform}:${accountId}`)
+  }
+
+  async getAccount(accountId, platform) {
+    if (!accountId || !platform) {
+      return null
+    }
+
+    const serviceMap = {
+      claude: require('./claudeAccountService'),
+      'claude-console': require('./claudeConsoleAccountService'),
+      gemini: require('./geminiAccountService'),
+      'gemini-api': require('./geminiApiAccountService'),
+      openai: require('./openaiAccountService'),
+      'openai-responses': require('./openaiResponsesAccountService'),
+      azure_openai: require('./azureOpenaiAccountService'),
+      bedrock: require('./bedrockAccountService'),
+      droid: require('./droidAccountService'),
+      ccr: require('./ccrAccountService')
+    }
+
+    const service = serviceMap[platform]
+    if (!service || typeof service.getAccount !== 'function') {
+      return null
+    }
+
+    const result = await service.getAccount(accountId)
+
+    // 处理不同服务返回格式的差异
+    // Bedrock/CCR/Droid 等服务返回 { success, data } 格式
+    if (result && typeof result === 'object' && 'success' in result && 'data' in result) {
+      return result.success ? result.data : null
+    }
+
+    return result
+  }
+
+  async getAllAccountsByPlatform(platform) {
+    if (!platform) {
+      return []
+    }
+
+    const serviceMap = {
+      claude: require('./claudeAccountService'),
+      'claude-console': require('./claudeConsoleAccountService'),
+      gemini: require('./geminiAccountService'),
+      'gemini-api': require('./geminiApiAccountService'),
+      openai: require('./openaiAccountService'),
+      'openai-responses': require('./openaiResponsesAccountService'),
+      azure_openai: require('./azureOpenaiAccountService'),
+      bedrock: require('./bedrockAccountService'),
+      droid: require('./droidAccountService'),
+      ccr: require('./ccrAccountService')
+    }
+
+    const service = serviceMap[platform]
+    if (!service) {
+      return []
+    }
+
+    // Bedrock 特殊：返回 { success, data }
+    if (platform === 'bedrock' && typeof service.getAllAccounts === 'function') {
+      const result = await service.getAllAccounts()
+      return result?.success ? result.data || [] : []
+    }
+
+    if (platform === 'openai-responses') {
+      return await service.getAllAccounts(true)
+    }
+
+    if (typeof service.getAllAccounts !== 'function') {
+      return []
+    }
+
+    return await service.getAllAccounts()
+  }
+
+  async _getAccountBalanceForAccount(account, platform, options = {}) {
+    const queryMode = this._parseQueryMode(options.queryApi)
+    const useCache = options.useCache !== false
+
+    const accountId = account?.id
+    if (!accountId) {
+      // 如果账户缺少 id，返回空响应而不是抛出错误，避免接口报错和UI错误
+      this.logger.warn('账户缺少 id，返回空余额数据', { account, platform })
+      return this._buildResponse(
+        {
+          status: 'error',
+          errorMessage: '账户数据异常',
+          balance: null,
+          currency: 'USD',
+          quota: null,
+          statistics: {},
+          lastRefreshAt: new Date().toISOString()
+        },
+        'unknown',
+        platform,
+        'local',
+        null,
+        { scriptEnabled: false, scriptConfigured: false }
+      )
+    }
+
+    // 余额脚本配置状态（用于前端控制"刷新余额"按钮）
+    let scriptConfig = null
+    let scriptConfigured = false
+    if (typeof this.redis?.getBalanceScriptConfig === 'function') {
+      scriptConfig = await this.redis.getBalanceScriptConfig(platform, accountId)
+      scriptConfigured = !!(
+        scriptConfig &&
+        scriptConfig.scriptBody &&
+        String(scriptConfig.scriptBody).trim().length > 0
+      )
+    }
+    const scriptEnabled = isBalanceScriptEnabled()
+    const scriptMeta = { scriptEnabled, scriptConfigured }
+
+    const localBalance = await this._getBalanceFromLocal(accountId, platform)
+    const localStatistics = localBalance.statistics || {}
+
+    const quotaFromLocal = this._buildQuotaFromLocal(account, localStatistics)
+
+    // 安全限制：queryApi=auto 仅用于 Antigravity（gemini + oauthProvider=antigravity）账户
+    const effectiveQueryMode =
+      queryMode === 'auto' && !(platform === 'gemini' && account?.oauthProvider === 'antigravity')
+        ? 'local'
+        : queryMode
+
+    // local: 仅本地统计/缓存；auto: 优先缓存，无缓存则尝试远程 Provider（并缓存结果）
+    if (effectiveQueryMode !== 'api') {
+      if (useCache) {
+        const cached = await this.redis.getAccountBalance(platform, accountId)
+        if (cached && cached.status === 'success') {
+          return this._buildResponse(
+            {
+              status: cached.status,
+              errorMessage: cached.errorMessage,
+              balance: quotaFromLocal.balance ?? cached.balance,
+              currency: quotaFromLocal.currency || cached.currency || 'USD',
+              quota: quotaFromLocal.quota || cached.quota || null,
+              statistics: localStatistics,
+              lastRefreshAt: cached.lastRefreshAt
+            },
+            accountId,
+            platform,
+            'cache',
+            cached.ttlSeconds,
+            scriptMeta
+          )
+        }
+      }
+
+      if (effectiveQueryMode === 'local') {
+        return this._buildResponse(
+          {
+            status: 'success',
+            errorMessage: null,
+            balance: quotaFromLocal.balance,
+            currency: quotaFromLocal.currency || 'USD',
+            quota: quotaFromLocal.quota,
+            statistics: localStatistics,
+            lastRefreshAt: localBalance.lastCalculated
+          },
+          accountId,
+          platform,
+          'local',
+          null,
+          scriptMeta
+        )
+      }
+    }
+
+    // 强制查询：优先脚本（如启用且已配置），否则调用 Provider；失败自动降级到本地统计
+    let providerResult
+
+    if (scriptEnabled && scriptConfigured) {
+      providerResult = await this._getBalanceFromScript(scriptConfig, accountId, platform)
+    } else {
+      const provider = this.providers.get(platform)
+      if (!provider) {
+        return this._buildResponse(
+          {
+            status: 'error',
+            errorMessage: `不支持的平台: ${platform}`,
+            balance: quotaFromLocal.balance,
+            currency: quotaFromLocal.currency || 'USD',
+            quota: quotaFromLocal.quota,
+            statistics: localStatistics,
+            lastRefreshAt: new Date().toISOString()
+          },
+          accountId,
+          platform,
+          'local',
+          null,
+          scriptMeta
+        )
+      }
+      providerResult = await this._getBalanceFromProvider(provider, account)
+    }
+
+    const isRemoteSuccess =
+      providerResult.status === 'success' && ['api', 'script'].includes(providerResult.queryMethod)
+
+    // 仅缓存“真实远程查询成功”的结果，避免把字段/本地降级结果当作 API 结果缓存 1h
+    if (isRemoteSuccess) {
+      await this.redis.setAccountBalance(
+        platform,
+        accountId,
+        providerResult,
+        this.CACHE_TTL_SECONDS
+      )
+    }
+
+    const source = isRemoteSuccess ? 'api' : 'local'
+
+    return this._buildResponse(
+      {
+        status: providerResult.status,
+        errorMessage: providerResult.errorMessage,
+        balance: quotaFromLocal.balance ?? providerResult.balance,
+        currency: quotaFromLocal.currency || providerResult.currency || 'USD',
+        quota: quotaFromLocal.quota || providerResult.quota || null,
+        statistics: localStatistics,
+        lastRefreshAt: providerResult.lastRefreshAt
+      },
+      accountId,
+      platform,
+      source,
+      null,
+      scriptMeta
+    )
+  }
+
+  async _getBalanceFromScript(scriptConfig, accountId, platform) {
+    try {
+      const result = await balanceScriptService.execute({
+        scriptBody: scriptConfig.scriptBody,
+        timeoutSeconds: scriptConfig.timeoutSeconds || 10,
+        variables: {
+          baseUrl: scriptConfig.baseUrl || '',
+          apiKey: scriptConfig.apiKey || '',
+          token: scriptConfig.token || '',
+          accountId,
+          platform,
+          extra: scriptConfig.extra || ''
+        }
+      })
+
+      const mapped = result?.mapped || {}
+      return {
+        status: mapped.status || 'error',
+        balance: typeof mapped.balance === 'number' ? mapped.balance : null,
+        currency: mapped.currency || 'USD',
+        quota: mapped.quota || null,
+        queryMethod: 'api',
+        rawData: mapped.rawData || result?.response?.data || null,
+        lastRefreshAt: new Date().toISOString(),
+        errorMessage: mapped.errorMessage || ''
+      }
+    } catch (error) {
+      return {
+        status: 'error',
+        balance: null,
+        currency: 'USD',
+        quota: null,
+        queryMethod: 'api',
+        rawData: null,
+        lastRefreshAt: new Date().toISOString(),
+        errorMessage: error.message || '脚本执行失败'
+      }
+    }
+  }
+
+  async _getBalanceFromProvider(provider, account) {
+    try {
+      const result = await provider.queryBalance(account)
+      return {
+        status: 'success',
+        balance: typeof result?.balance === 'number' ? result.balance : null,
+        currency: result?.currency || 'USD',
+        quota: result?.quota || null,
+        queryMethod: result?.queryMethod || 'api',
+        rawData: result?.rawData || null,
+        lastRefreshAt: new Date().toISOString(),
+        errorMessage: ''
+      }
+    } catch (error) {
+      return {
+        status: 'error',
+        balance: null,
+        currency: 'USD',
+        quota: null,
+        queryMethod: 'api',
+        rawData: null,
+        lastRefreshAt: new Date().toISOString(),
+        errorMessage: error.message || '查询失败'
+      }
+    }
+  }
+
+  async _getBalanceFromLocal(accountId, platform) {
+    const cached = await this.redis.getLocalBalance(platform, accountId)
+    if (cached && cached.statistics) {
+      return cached
+    }
+
+    const statistics = await this._computeLocalStatistics(accountId)
+    const localBalance = {
+      status: 'success',
+      balance: null,
+      currency: 'USD',
+      statistics,
+      queryMethod: 'local',
+      lastCalculated: new Date().toISOString()
+    }
+
+    await this.redis.setLocalBalance(platform, accountId, localBalance, this.LOCAL_TTL_SECONDS)
+    return localBalance
+  }
+
+  async _computeLocalStatistics(accountId) {
+    const safeNumber = (value) => {
+      const num = Number(value)
+      return Number.isFinite(num) ? num : 0
+    }
+
+    try {
+      const usageStats = await this.redis.getAccountUsageStats(accountId)
+      const dailyCost = safeNumber(usageStats?.daily?.cost || 0)
+      const monthlyCost = await this._computeMonthlyCost(accountId)
+      const totalCost = await this._computeTotalCost(accountId)
+
+      return {
+        totalCost,
+        dailyCost,
+        monthlyCost,
+        totalRequests: safeNumber(usageStats?.total?.requests || 0),
+        dailyRequests: safeNumber(usageStats?.daily?.requests || 0),
+        monthlyRequests: safeNumber(usageStats?.monthly?.requests || 0)
+      }
+    } catch (error) {
+      this.logger.debug(`本地统计计算失败: ${accountId}`, error)
+      return {
+        totalCost: 0,
+        dailyCost: 0,
+        monthlyCost: 0,
+        totalRequests: 0,
+        dailyRequests: 0,
+        monthlyRequests: 0
+      }
+    }
+  }
+
+  async _computeMonthlyCost(accountId) {
+    const tzDate = this.redis.getDateInTimezone(new Date())
+    const currentMonth = `${tzDate.getUTCFullYear()}-${String(tzDate.getUTCMonth() + 1).padStart(
+      2,
+      '0'
+    )}`
+
+    const pattern = `account_usage:model:monthly:${accountId}:*:${currentMonth}`
+    return await this._sumModelCostsByKeysPattern(pattern)
+  }
+
+  async _computeTotalCost(accountId) {
+    const pattern = `account_usage:model:monthly:${accountId}:*:*`
+    return await this._sumModelCostsByKeysPattern(pattern)
+  }
+
+  async _sumModelCostsByKeysPattern(pattern) {
+    try {
+      const client = this.redis.getClientSafe()
+      let totalCost = 0
+      let cursor = '0'
+      const scanCount = 200
+      let iterations = 0
+      const maxIterations = 2000
+
+      do {
+        const [nextCursor, keys] = await client.scan(cursor, 'MATCH', pattern, 'COUNT', scanCount)
+        cursor = nextCursor
+        iterations += 1
+
+        if (!keys || keys.length === 0) {
+          continue
+        }
+
+        const pipeline = client.pipeline()
+        keys.forEach((key) => pipeline.hgetall(key))
+        const results = await pipeline.exec()
+
+        for (let i = 0; i < results.length; i += 1) {
+          const [, data] = results[i] || []
+          if (!data || Object.keys(data).length === 0) {
+            continue
+          }
+
+          const parts = String(keys[i]).split(':')
+          const model = parts[4] || 'unknown'
+
+          const usage = {
+            input_tokens: parseInt(data.inputTokens || 0),
+            output_tokens: parseInt(data.outputTokens || 0),
+            cache_creation_input_tokens: parseInt(data.cacheCreateTokens || 0),
+            cache_read_input_tokens: parseInt(data.cacheReadTokens || 0)
+          }
+
+          const costResult = CostCalculator.calculateCost(usage, model)
+          totalCost += costResult.costs.total || 0
+        }
+
+        if (iterations >= maxIterations) {
+          this.logger.warn(`SCAN 次数超过上限，停止汇总：${pattern}`)
+          break
+        }
+      } while (cursor !== '0')
+
+      return totalCost
+    } catch (error) {
+      this.logger.debug(`汇总模型费用失败: ${pattern}`, error)
+      return 0
+    }
+  }
+
+  _buildQuotaFromLocal(account, statistics) {
+    if (!account || !Object.prototype.hasOwnProperty.call(account, 'dailyQuota')) {
+      return { balance: null, currency: null, quota: null }
+    }
+
+    const dailyQuota = Number(account.dailyQuota || 0)
+    const used = Number(statistics?.dailyCost || 0)
+
+    const resetAt = this._computeNextResetAt(account.quotaResetTime || '00:00')
+
+    // 不限制
+    if (!Number.isFinite(dailyQuota) || dailyQuota <= 0) {
+      return {
+        balance: null,
+        currency: 'USD',
+        quota: {
+          daily: Infinity,
+          used,
+          remaining: Infinity,
+          percentage: 0,
+          unlimited: true,
+          resetAt
+        }
+      }
+    }
+
+    const remaining = Math.max(0, dailyQuota - used)
+    const percentage = dailyQuota > 0 ? (used / dailyQuota) * 100 : 0
+
+    return {
+      balance: remaining,
+      currency: 'USD',
+      quota: {
+        daily: dailyQuota,
+        used,
+        remaining,
+        resetAt,
+        percentage: Math.round(percentage * 100) / 100
+      }
+    }
+  }
+
+  _computeNextResetAt(resetTime) {
+    const now = new Date()
+    const tzNow = this.redis.getDateInTimezone(now)
+    const offsetMs = tzNow.getTime() - now.getTime()
+
+    const [h, m] = String(resetTime || '00:00')
+      .split(':')
+      .map((n) => parseInt(n, 10))
+
+    const resetHour = Number.isFinite(h) ? h : 0
+    const resetMinute = Number.isFinite(m) ? m : 0
+
+    const year = tzNow.getUTCFullYear()
+    const month = tzNow.getUTCMonth()
+    const day = tzNow.getUTCDate()
+
+    let resetAtMs = Date.UTC(year, month, day, resetHour, resetMinute, 0, 0) - offsetMs
+    if (resetAtMs <= now.getTime()) {
+      resetAtMs += 24 * 60 * 60 * 1000
+    }
+
+    return new Date(resetAtMs).toISOString()
+  }
+
+  _buildResponse(balanceData, accountId, platform, source, ttlSeconds = null, extraData = {}) {
+    const now = new Date()
+
+    const amount = typeof balanceData.balance === 'number' ? balanceData.balance : null
+    const currency = balanceData.currency || 'USD'
+
+    let cacheExpiresAt = null
+    if (source === 'cache') {
+      const ttl =
+        typeof ttlSeconds === 'number' && ttlSeconds > 0 ? ttlSeconds : this.CACHE_TTL_SECONDS
+      cacheExpiresAt = new Date(Date.now() + ttl * 1000).toISOString()
+    }
+
+    return {
+      success: true,
+      data: {
+        accountId,
+        platform,
+        balance:
+          typeof amount === 'number'
+            ? {
+                amount,
+                currency,
+                formattedAmount: this._formatCurrency(amount, currency)
+              }
+            : null,
+        quota: balanceData.quota || null,
+        statistics: balanceData.statistics || {},
+        source,
+        lastRefreshAt: balanceData.lastRefreshAt || now.toISOString(),
+        cacheExpiresAt,
+        status: balanceData.status || 'success',
+        error: balanceData.errorMessage || null,
+        ...(extraData && typeof extraData === 'object' ? extraData : {})
+      }
+    }
+  }
+
+  _formatCurrency(amount, currency = 'USD') {
+    try {
+      if (typeof amount !== 'number' || !Number.isFinite(amount)) {
+        return 'N/A'
+      }
+      return new Intl.NumberFormat('en-US', { style: 'currency', currency }).format(amount)
+    } catch (error) {
+      return `$${amount.toFixed(2)}`
+    }
+  }
+
+  _parseBoolean(value) {
+    if (typeof value === 'boolean') {
+      return value
+    }
+    if (typeof value !== 'string') {
+      return null
+    }
+    const normalized = value.trim().toLowerCase()
+    if (normalized === 'true' || normalized === '1' || normalized === 'yes') {
+      return true
+    }
+    if (normalized === 'false' || normalized === '0' || normalized === 'no') {
+      return false
+    }
+    return null
+  }
+
+  _parseQueryMode(value) {
+    if (value === 'auto') {
+      return 'auto'
+    }
+    const parsed = this._parseBoolean(value)
+    return parsed ? 'api' : 'local'
+  }
+
+  async _mapWithConcurrency(items, limit, mapper) {
+    const concurrency = Math.max(1, Number(limit) || 1)
+    const list = Array.isArray(items) ? items : []
+
+    const results = new Array(list.length)
+    let nextIndex = 0
+
+    const workers = new Array(Math.min(concurrency, list.length)).fill(null).map(async () => {
+      while (nextIndex < list.length) {
+        const currentIndex = nextIndex
+        nextIndex += 1
+        results[currentIndex] = await mapper(list[currentIndex], currentIndex)
+      }
+    })
+
+    await Promise.all(workers)
+    return results
+  }
+}
+
+const accountBalanceService = new AccountBalanceService()
+module.exports = accountBalanceService
+module.exports.AccountBalanceService = AccountBalanceService

src/services/accountTestSchedulerService.js

@@ -0,0 +1,420 @@
+/**
+ * 账户定时测试调度服务
+ * 使用 node-cron 支持 crontab 表达式，为每个账户创建独立的定时任务
+ */
+
+const cron = require('node-cron')
+const redis = require('../models/redis')
+const logger = require('../utils/logger')
+
+class AccountTestSchedulerService {
+  constructor() {
+    // 存储每个账户的 cron 任务: Map<string, { task: ScheduledTask, cronExpression: string }>
+    this.scheduledTasks = new Map()
+    // 定期刷新配置的间隔 (毫秒)
+    this.refreshIntervalMs = 60 * 1000
+    this.refreshInterval = null
+    // 当前正在测试的账户
+    this.testingAccounts = new Set()
+    // 是否已启动
+    this.isStarted = false
+  }
+
+  /**
+   * 验证 cron 表达式是否有效
+   * @param {string} cronExpression - cron 表达式
+   * @returns {boolean}
+   */
+  validateCronExpression(cronExpression) {
+    // 长度检查（防止 DoS）
+    if (!cronExpression || cronExpression.length > 100) {
+      return false
+    }
+    return cron.validate(cronExpression)
+  }
+
+  /**
+   * 启动调度器
+   */
+  async start() {
+    if (this.isStarted) {
+      logger.warn('⚠️ Account test scheduler is already running')
+      return
+    }
+
+    this.isStarted = true
+    logger.info('🚀 Starting account test scheduler service (node-cron mode)')
+
+    // 初始化所有已配置账户的定时任务
+    await this._refreshAllTasks()
+
+    // 定期刷新配置，以便动态添加/修改的配置能生效
+    this.refreshInterval = setInterval(() => {
+      this._refreshAllTasks()
+    }, this.refreshIntervalMs)
+
+    logger.info(
+      `📅 Account test scheduler started (refreshing configs every ${this.refreshIntervalMs / 1000}s)`
+    )
+  }
+
+  /**
+   * 停止调度器
+   */
+  stop() {
+    if (this.refreshInterval) {
+      clearInterval(this.refreshInterval)
+      this.refreshInterval = null
+    }
+
+    // 停止所有 cron 任务
+    for (const [accountKey, taskInfo] of this.scheduledTasks.entries()) {
+      taskInfo.task.stop()
+      logger.debug(`🛑 Stopped cron task for ${accountKey}`)
+    }
+    this.scheduledTasks.clear()
+
+    this.isStarted = false
+    logger.info('🛑 Account test scheduler stopped')
+  }
+
+  /**
+   * 刷新所有账户的定时任务
+   * @private
+   */
+  async _refreshAllTasks() {
+    try {
+      const platforms = ['claude', 'gemini', 'openai']
+      const activeAccountKeys = new Set()
+
+      // 并行加载所有平台的配置
+      const allEnabledAccounts = await Promise.all(
+        platforms.map((platform) =>
+          redis
+            .getEnabledTestAccounts(platform)
+            .then((accounts) => accounts.map((acc) => ({ ...acc, platform })))
+            .catch((error) => {
+              logger.warn(`⚠️ Failed to load test accounts for platform ${platform}:`, error)
+              return []
+            })
+        )
+      )
+
+      // 展平平台数据
+      const flatAccounts = allEnabledAccounts.flat()
+
+      for (const { accountId, cronExpression, model, platform } of flatAccounts) {
+        if (!cronExpression) {
+          logger.warn(
+            `⚠️ Account ${accountId} (${platform}) has no valid cron expression, skipping`
+          )
+          continue
+        }
+
+        const accountKey = `${platform}:${accountId}`
+        activeAccountKeys.add(accountKey)
+
+        // 检查是否需要更新任务
+        const existingTask = this.scheduledTasks.get(accountKey)
+        if (existingTask) {
+          // 如果 cron 表达式和模型都没变，不需要更新
+          if (existingTask.cronExpression === cronExpression && existingTask.model === model) {
+            continue
+          }
+          // 配置变了，停止旧任务
+          existingTask.task.stop()
+          logger.info(`🔄 Updating cron task for ${accountKey}: ${cronExpression}, model: ${model}`)
+        } else {
+          logger.info(`➕ Creating cron task for ${accountKey}: ${cronExpression}, model: ${model}`)
+        }
+
+        // 创建新的 cron 任务
+        this._createCronTask(accountId, platform, cronExpression, model)
+      }
+
+      // 清理已删除或禁用的账户任务
+      for (const [accountKey, taskInfo] of this.scheduledTasks.entries()) {
+        if (!activeAccountKeys.has(accountKey)) {
+          taskInfo.task.stop()
+          this.scheduledTasks.delete(accountKey)
+          logger.info(`➖ Removed cron task for ${accountKey} (disabled or deleted)`)
+        }
+      }
+    } catch (error) {
+      logger.error('❌ Error refreshing account test tasks:', error)
+    }
+  }
+
+  /**
+   * 为单个账户创建 cron 任务
+   * @param {string} accountId
+   * @param {string} platform
+   * @param {string} cronExpression
+   * @param {string} model - 测试使用的模型
+   * @private
+   */
+  _createCronTask(accountId, platform, cronExpression, model) {
+    const accountKey = `${platform}:${accountId}`
+
+    // 验证 cron 表达式
+    if (!this.validateCronExpression(cronExpression)) {
+      logger.error(`❌ Invalid cron expression for ${accountKey}: ${cronExpression}`)
+      return
+    }
+
+    const task = cron.schedule(
+      cronExpression,
+      async () => {
+        await this._runAccountTest(accountId, platform, model)
+      },
+      {
+        scheduled: true,
+        timezone: process.env.TZ || 'Asia/Shanghai'
+      }
+    )
+
+    this.scheduledTasks.set(accountKey, {
+      task,
+      cronExpression,
+      model,
+      accountId,
+      platform
+    })
+  }
+
+  /**
+   * 执行单个账户测试
+   * @param {string} accountId - 账户ID
+   * @param {string} platform - 平台类型
+   * @param {string} model - 测试使用的模型
+   * @private
+   */
+  async _runAccountTest(accountId, platform, model) {
+    const accountKey = `${platform}:${accountId}`
+
+    // 避免重复测试
+    if (this.testingAccounts.has(accountKey)) {
+      logger.debug(`⏳ Account ${accountKey} is already being tested, skipping`)
+      return
+    }
+
+    this.testingAccounts.add(accountKey)
+
+    try {
+      logger.info(
+        `🧪 Running scheduled test for ${platform} account: ${accountId} (model: ${model})`
+      )
+
+      let testResult
+
+      // 根据平台调用对应的测试方法
+      switch (platform) {
+        case 'claude':
+          testResult = await this._testClaudeAccount(accountId, model)
+          break
+        case 'gemini':
+          testResult = await this._testGeminiAccount(accountId, model)
+          break
+        case 'openai':
+          testResult = await this._testOpenAIAccount(accountId, model)
+          break
+        default:
+          testResult = {
+            success: false,
+            error: `Unsupported platform: ${platform}`,
+            timestamp: new Date().toISOString()
+          }
+      }
+
+      // 保存测试结果
+      await redis.saveAccountTestResult(accountId, platform, testResult)
+
+      // 更新最后测试时间
+      await redis.setAccountLastTestTime(accountId, platform)
+
+      // 记录日志
+      if (testResult.success) {
+        logger.info(
+          `✅ Scheduled test passed for ${platform} account ${accountId} (${testResult.latencyMs}ms)`
+        )
+      } else {
+        logger.warn(
+          `❌ Scheduled test failed for ${platform} account ${accountId}: ${testResult.error}`
+        )
+      }
+
+      return testResult
+    } catch (error) {
+      logger.error(`❌ Error testing ${platform} account ${accountId}:`, error)
+
+      const errorResult = {
+        success: false,
+        error: error.message,
+        timestamp: new Date().toISOString()
+      }
+
+      await redis.saveAccountTestResult(accountId, platform, errorResult)
+      await redis.setAccountLastTestTime(accountId, platform)
+
+      return errorResult
+    } finally {
+      this.testingAccounts.delete(accountKey)
+    }
+  }
+
+  /**
+   * 测试 Claude 账户
+   * @param {string} accountId
+   * @param {string} model - 测试使用的模型
+   * @private
+   */
+  async _testClaudeAccount(accountId, model) {
+    const claudeRelayService = require('./claudeRelayService')
+    return await claudeRelayService.testAccountConnectionSync(accountId, model)
+  }
+
+  /**
+   * 测试 Gemini 账户
+   * @param {string} _accountId
+   * @param {string} _model
+   * @private
+   */
+  async _testGeminiAccount(_accountId, _model) {
+    // Gemini 测试暂时返回未实现
+    return {
+      success: false,
+      error: 'Gemini scheduled test not implemented yet',
+      timestamp: new Date().toISOString()
+    }
+  }
+
+  /**
+   * 测试 OpenAI 账户
+   * @param {string} _accountId
+   * @param {string} _model
+   * @private
+   */
+  async _testOpenAIAccount(_accountId, _model) {
+    // OpenAI 测试暂时返回未实现
+    return {
+      success: false,
+      error: 'OpenAI scheduled test not implemented yet',
+      timestamp: new Date().toISOString()
+    }
+  }
+
+  /**
+   * 手动触发账户测试
+   * @param {string} accountId - 账户ID
+   * @param {string} platform - 平台类型
+   * @param {string} model - 测试使用的模型
+   * @returns {Promise<Object>} 测试结果
+   */
+  async triggerTest(accountId, platform, model = 'claude-sonnet-4-5-20250929') {
+    logger.info(`🎯 Manual test triggered for ${platform} account: ${accountId} (model: ${model})`)
+    return await this._runAccountTest(accountId, platform, model)
+  }
+
+  /**
+   * 获取账户测试历史
+   * @param {string} accountId - 账户ID
+   * @param {string} platform - 平台类型
+   * @returns {Promise<Array>} 测试历史
+   */
+  async getTestHistory(accountId, platform) {
+    return await redis.getAccountTestHistory(accountId, platform)
+  }
+
+  /**
+   * 获取账户测试配置
+   * @param {string} accountId - 账户ID
+   * @param {string} platform - 平台类型
+   * @returns {Promise<Object|null>}
+   */
+  async getTestConfig(accountId, platform) {
+    return await redis.getAccountTestConfig(accountId, platform)
+  }
+
+  /**
+   * 设置账户测试配置
+   * @param {string} accountId - 账户ID
+   * @param {string} platform - 平台类型
+   * @param {Object} testConfig - 测试配置 { enabled: boolean, cronExpression: string, model: string }
+   * @returns {Promise<void>}
+   */
+  async setTestConfig(accountId, platform, testConfig) {
+    // 验证 cron 表达式
+    if (testConfig.cronExpression && !this.validateCronExpression(testConfig.cronExpression)) {
+      throw new Error(`Invalid cron expression: ${testConfig.cronExpression}`)
+    }
+
+    await redis.saveAccountTestConfig(accountId, platform, testConfig)
+    logger.info(
+      `📝 Test config updated for ${platform} account ${accountId}: enabled=${testConfig.enabled}, cronExpression=${testConfig.cronExpression}, model=${testConfig.model}`
+    )
+
+    // 立即刷新任务，使配置立即生效
+    if (this.isStarted) {
+      await this._refreshAllTasks()
+    }
+  }
+
+  /**
+   * 更新单个账户的定时任务（配置变更时调用）
+   * @param {string} accountId
+   * @param {string} platform
+   */
+  async refreshAccountTask(accountId, platform) {
+    if (!this.isStarted) {
+      return
+    }
+
+    const accountKey = `${platform}:${accountId}`
+    const testConfig = await redis.getAccountTestConfig(accountId, platform)
+
+    // 停止现有任务
+    const existingTask = this.scheduledTasks.get(accountKey)
+    if (existingTask) {
+      existingTask.task.stop()
+      this.scheduledTasks.delete(accountKey)
+    }
+
+    // 如果启用且有有效的 cron 表达式，创建新任务
+    if (testConfig?.enabled && testConfig?.cronExpression) {
+      this._createCronTask(accountId, platform, testConfig.cronExpression, testConfig.model)
+      logger.info(
+        `🔄 Refreshed cron task for ${accountKey}: ${testConfig.cronExpression}, model: ${testConfig.model}`
+      )
+    }
+  }
+
+  /**
+   * 获取调度器状态
+   * @returns {Object}
+   */
+  getStatus() {
+    const tasks = []
+    for (const [accountKey, taskInfo] of this.scheduledTasks.entries()) {
+      tasks.push({
+        accountKey,
+        accountId: taskInfo.accountId,
+        platform: taskInfo.platform,
+        cronExpression: taskInfo.cronExpression,
+        model: taskInfo.model
+      })
+    }
+
+    return {
+      running: this.isStarted,
+      refreshIntervalMs: this.refreshIntervalMs,
+      scheduledTasksCount: this.scheduledTasks.size,
+      scheduledTasks: tasks,
+      currentlyTesting: Array.from(this.testingAccounts)
+    }
+  }
+}
+
+// 单例模式
+const accountTestSchedulerService = new AccountTestSchedulerService()
+
+module.exports = accountTestSchedulerService

src/services/antigravityClient.js

@@ -0,0 +1,595 @@
+const axios = require('axios')
+const https = require('https')
+const { v4: uuidv4 } = require('uuid')
+
+const ProxyHelper = require('../utils/proxyHelper')
+const logger = require('../utils/logger')
+const {
+  mapAntigravityUpstreamModel,
+  normalizeAntigravityModelInput,
+  getAntigravityModelMetadata
+} = require('../utils/antigravityModel')
+const { cleanJsonSchemaForGemini } = require('../utils/geminiSchemaCleaner')
+const { dumpAntigravityUpstreamRequest } = require('../utils/antigravityUpstreamDump')
+
+const keepAliveAgent = new https.Agent({
+  keepAlive: true,
+  keepAliveMsecs: 30000,
+  timeout: 120000,
+  maxSockets: 100,
+  maxFreeSockets: 10
+})
+
+function getAntigravityApiUrl() {
+  return process.env.ANTIGRAVITY_API_URL || 'https://daily-cloudcode-pa.sandbox.googleapis.com'
+}
+
+function normalizeBaseUrl(url) {
+  const str = String(url || '').trim()
+  return str.endsWith('/') ? str.slice(0, -1) : str
+}
+
+function getAntigravityApiUrlCandidates() {
+  const configured = normalizeBaseUrl(getAntigravityApiUrl())
+  const daily = 'https://daily-cloudcode-pa.sandbox.googleapis.com'
+  const prod = 'https://cloudcode-pa.googleapis.com'
+
+  // 若显式配置了自定义 base url，则只使用该地址（不做 fallback，避免意外路由到别的环境）。
+  if (process.env.ANTIGRAVITY_API_URL) {
+    return [configured]
+  }
+
+  // 默认行为：优先 daily（与旧逻辑一致），失败时再尝试 prod（对齐 CLIProxyAPI）。
+  if (configured === normalizeBaseUrl(daily)) {
+    return [configured, prod]
+  }
+  if (configured === normalizeBaseUrl(prod)) {
+    return [configured, daily]
+  }
+
+  return [configured, prod, daily].filter(Boolean)
+}
+
+function getAntigravityHeaders(accessToken, baseUrl) {
+  const resolvedBaseUrl = baseUrl || getAntigravityApiUrl()
+  let host = 'daily-cloudcode-pa.sandbox.googleapis.com'
+  try {
+    host = new URL(resolvedBaseUrl).host || host
+  } catch (e) {
+    // ignore
+  }
+
+  return {
+    Host: host,
+    'User-Agent': process.env.ANTIGRAVITY_USER_AGENT || 'antigravity/1.11.3 windows/amd64',
+    Authorization: `Bearer ${accessToken}`,
+    'Content-Type': 'application/json',
+    'Accept-Encoding': 'gzip',
+    requestType: 'agent'
+  }
+}
+
+function generateAntigravityProjectId() {
+  return `ag-${uuidv4().replace(/-/g, '').slice(0, 16)}`
+}
+
+function generateAntigravitySessionId() {
+  return `sess-${uuidv4()}`
+}
+
+function resolveAntigravityProjectId(projectId, requestData) {
+  const candidate = projectId || requestData?.project || requestData?.projectId || null
+  return candidate || generateAntigravityProjectId()
+}
+
+function resolveAntigravitySessionId(sessionId, requestData) {
+  const candidate =
+    sessionId || requestData?.request?.sessionId || requestData?.request?.session_id || null
+  return candidate || generateAntigravitySessionId()
+}
+
+function buildAntigravityEnvelope({ requestData, projectId, sessionId, userPromptId }) {
+  const model = mapAntigravityUpstreamModel(requestData?.model)
+  const resolvedProjectId = resolveAntigravityProjectId(projectId, requestData)
+  const resolvedSessionId = resolveAntigravitySessionId(sessionId, requestData)
+  const requestPayload = {
+    ...(requestData?.request || {})
+  }
+
+  if (requestPayload.session_id !== undefined) {
+    delete requestPayload.session_id
+  }
+  requestPayload.sessionId = resolvedSessionId
+
+  const envelope = {
+    project: resolvedProjectId,
+    requestId: `req-${uuidv4()}`,
+    model,
+    userAgent: 'antigravity',
+    request: {
+      ...requestPayload
+    }
+  }
+
+  if (userPromptId) {
+    envelope.user_prompt_id = userPromptId
+    envelope.userPromptId = userPromptId
+  }
+
+  normalizeAntigravityEnvelope(envelope)
+  return { model, envelope }
+}
+
+function normalizeAntigravityThinking(model, requestPayload) {
+  if (!requestPayload || typeof requestPayload !== 'object') {
+    return
+  }
+
+  const { generationConfig } = requestPayload
+  if (!generationConfig || typeof generationConfig !== 'object') {
+    return
+  }
+  const { thinkingConfig } = generationConfig
+  if (!thinkingConfig || typeof thinkingConfig !== 'object') {
+    return
+  }
+
+  const normalizedModel = normalizeAntigravityModelInput(model)
+  if (thinkingConfig.thinkingLevel && !normalizedModel.startsWith('gemini-3-')) {
+    delete thinkingConfig.thinkingLevel
+  }
+
+  const metadata = getAntigravityModelMetadata(normalizedModel)
+  if (metadata && !metadata.thinking) {
+    delete generationConfig.thinkingConfig
+    return
+  }
+  if (!metadata || !metadata.thinking) {
+    return
+  }
+
+  const budgetRaw = Number(thinkingConfig.thinkingBudget)
+  if (!Number.isFinite(budgetRaw)) {
+    return
+  }
+  let budget = Math.trunc(budgetRaw)
+
+  const minBudget = Number.isFinite(metadata.thinking.min) ? metadata.thinking.min : null
+  const maxBudget = Number.isFinite(metadata.thinking.max) ? metadata.thinking.max : null
+
+  if (maxBudget !== null && budget > maxBudget) {
+    budget = maxBudget
+  }
+
+  let effectiveMax = Number.isFinite(generationConfig.maxOutputTokens)
+    ? generationConfig.maxOutputTokens
+    : null
+  let setDefaultMax = false
+  if (!effectiveMax && metadata.maxCompletionTokens) {
+    effectiveMax = metadata.maxCompletionTokens
+    setDefaultMax = true
+  }
+
+  if (effectiveMax && budget >= effectiveMax) {
+    budget = Math.max(0, effectiveMax - 1)
+  }
+
+  if (minBudget !== null && budget >= 0 && budget < minBudget) {
+    delete generationConfig.thinkingConfig
+    return
+  }
+
+  thinkingConfig.thinkingBudget = budget
+  if (setDefaultMax) {
+    generationConfig.maxOutputTokens = effectiveMax
+  }
+}
+
+function normalizeAntigravityEnvelope(envelope) {
+  if (!envelope || typeof envelope !== 'object') {
+    return
+  }
+  const model = String(envelope.model || '')
+  const requestPayload = envelope.request
+  if (!requestPayload || typeof requestPayload !== 'object') {
+    return
+  }
+
+  if (requestPayload.safetySettings !== undefined) {
+    delete requestPayload.safetySettings
+  }
+
+  // 对齐 CLIProxyAPI：有 tools 时默认启用 VALIDATED（除非显式 NONE）
+  if (Array.isArray(requestPayload.tools) && requestPayload.tools.length > 0) {
+    const existing = requestPayload?.toolConfig?.functionCallingConfig || null
+    if (existing?.mode !== 'NONE') {
+      const nextCfg = { ...(existing || {}), mode: 'VALIDATED' }
+      requestPayload.toolConfig = { functionCallingConfig: nextCfg }
+    }
+  }
+
+  // 对齐 CLIProxyAPI：非 Claude 模型移除 maxOutputTokens（Antigravity 环境不稳定）
+  normalizeAntigravityThinking(model, requestPayload)
+  if (!model.includes('claude')) {
+    if (requestPayload.generationConfig && typeof requestPayload.generationConfig === 'object') {
+      delete requestPayload.generationConfig.maxOutputTokens
+    }
+    return
+  }
+
+  // Claude 模型：parametersJsonSchema -> parameters + schema 清洗（避免 $schema / additionalProperties 等触发 400）
+  if (!Array.isArray(requestPayload.tools)) {
+    return
+  }
+
+  for (const tool of requestPayload.tools) {
+    if (!tool || typeof tool !== 'object') {
+      continue
+    }
+    const decls = Array.isArray(tool.functionDeclarations)
+      ? tool.functionDeclarations
+      : Array.isArray(tool.function_declarations)
+        ? tool.function_declarations
+        : null
+
+    if (!decls) {
+      continue
+    }
+
+    for (const decl of decls) {
+      if (!decl || typeof decl !== 'object') {
+        continue
+      }
+      let schema =
+        decl.parametersJsonSchema !== undefined ? decl.parametersJsonSchema : decl.parameters
+      if (typeof schema === 'string' && schema) {
+        try {
+          schema = JSON.parse(schema)
+        } catch (_) {
+          schema = null
+        }
+      }
+
+      decl.parameters = cleanJsonSchemaForGemini(schema)
+      delete decl.parametersJsonSchema
+    }
+  }
+}
+
+async function request({
+  accessToken,
+  proxyConfig = null,
+  requestData,
+  projectId = null,
+  sessionId = null,
+  userPromptId = null,
+  stream = false,
+  signal = null,
+  params = null,
+  timeoutMs = null
+}) {
+  const { model, envelope } = buildAntigravityEnvelope({
+    requestData,
+    projectId,
+    sessionId,
+    userPromptId
+  })
+
+  const proxyAgent = ProxyHelper.createProxyAgent(proxyConfig)
+  let endpoints = getAntigravityApiUrlCandidates()
+
+  // Claude 模型在 sandbox(daily) 环境下对 tool_use/tool_result 的兼容性不稳定，优先走 prod。
+  // 保持可配置优先：若用户显式设置了 ANTIGRAVITY_API_URL，则不改变顺序。
+  if (!process.env.ANTIGRAVITY_API_URL && String(model).includes('claude')) {
+    const prodHost = 'cloudcode-pa.googleapis.com'
+    const dailyHost = 'daily-cloudcode-pa.sandbox.googleapis.com'
+    const ordered = []
+    for (const u of endpoints) {
+      if (String(u).includes(prodHost)) {
+        ordered.push(u)
+      }
+    }
+    for (const u of endpoints) {
+      if (!String(u).includes(prodHost)) {
+        ordered.push(u)
+      }
+    }
+    // 去重并保持 prod -> daily 的稳定顺序
+    endpoints = Array.from(new Set(ordered)).sort((a, b) => {
+      const av = String(a)
+      const bv = String(b)
+      const aScore = av.includes(prodHost) ? 0 : av.includes(dailyHost) ? 1 : 2
+      const bScore = bv.includes(prodHost) ? 0 : bv.includes(dailyHost) ? 1 : 2
+      return aScore - bScore
+    })
+  }
+
+  const isRetryable = (error) => {
+    // 处理网络层面的连接重置或超时（常见于长请求被中间节点切断）
+    if (error.code === 'ECONNRESET' || error.code === 'ETIMEDOUT') {
+      return true
+    }
+
+    const status = error?.response?.status
+    if (status === 429) {
+      return true
+    }
+
+    // 400/404 的 “model unavailable / not found” 在不同环境间可能表现不同，允许 fallback。
+    if (status === 400 || status === 404) {
+      const data = error?.response?.data
+      const safeToString = (value) => {
+        if (typeof value === 'string') {
+          return value
+        }
+        if (value === null || value === undefined) {
+          return ''
+        }
+        // axios responseType=stream 时，data 可能是 stream（存在循环引用），不能 JSON.stringify
+        if (typeof value === 'object' && typeof value.pipe === 'function') {
+          return ''
+        }
+        if (Buffer.isBuffer(value)) {
+          try {
+            return value.toString('utf8')
+          } catch (_) {
+            return ''
+          }
+        }
+        if (typeof value === 'object') {
+          try {
+            return JSON.stringify(value)
+          } catch (_) {
+            return ''
+          }
+        }
+        return String(value)
+      }
+
+      const text = safeToString(data)
+      const msg = (text || '').toLowerCase()
+      return (
+        msg.includes('requested model is currently unavailable') ||
+        msg.includes('tool_use') ||
+        msg.includes('tool_result') ||
+        msg.includes('requested entity was not found') ||
+        msg.includes('not found')
+      )
+    }
+
+    return false
+  }
+
+  let lastError = null
+  let retriedAfterDelay = false
+
+  const attemptRequest = async () => {
+    for (let index = 0; index < endpoints.length; index += 1) {
+      const baseUrl = endpoints[index]
+      const url = `${baseUrl}/v1internal:${stream ? 'streamGenerateContent' : 'generateContent'}`
+
+      const axiosConfig = {
+        url,
+        method: 'POST',
+        ...(params ? { params } : {}),
+        headers: getAntigravityHeaders(accessToken, baseUrl),
+        data: envelope,
+        timeout: stream ? 0 : timeoutMs || 600000,
+        ...(stream ? { responseType: 'stream' } : {})
+      }
+
+      if (proxyAgent) {
+        axiosConfig.httpsAgent = proxyAgent
+        axiosConfig.proxy = false
+        if (index === 0) {
+          logger.info(
+            `🌐 Using proxy for Antigravity ${stream ? 'streamGenerateContent' : 'generateContent'}: ${ProxyHelper.getProxyDescription(proxyConfig)}`
+          )
+        }
+      } else {
+        axiosConfig.httpsAgent = keepAliveAgent
+      }
+
+      if (signal) {
+        axiosConfig.signal = signal
+      }
+
+      try {
+        dumpAntigravityUpstreamRequest({
+          requestId: envelope.requestId,
+          model,
+          stream,
+          url,
+          baseUrl,
+          params: axiosConfig.params || null,
+          headers: axiosConfig.headers,
+          envelope
+        }).catch(() => {})
+        const response = await axios(axiosConfig)
+        return { model, response }
+      } catch (error) {
+        lastError = error
+        const status = error?.response?.status || null
+
+        const hasNext = index + 1 < endpoints.length
+        if (hasNext && isRetryable(error)) {
+          logger.warn('⚠️ Antigravity upstream error, retrying with fallback baseUrl', {
+            status,
+            from: baseUrl,
+            to: endpoints[index + 1],
+            model
+          })
+          continue
+        }
+        throw error
+      }
+    }
+
+    throw lastError || new Error('Antigravity request failed')
+  }
+
+  try {
+    return await attemptRequest()
+  } catch (error) {
+    // 如果是 429 RESOURCE_EXHAUSTED 且尚未重试过，等待 2 秒后重试一次
+    const status = error?.response?.status
+    if (status === 429 && !retriedAfterDelay && !signal?.aborted) {
+      const data = error?.response?.data
+
+      // 安全地将 data 转为字符串，避免 stream 对象导致循环引用崩溃
+      const safeDataToString = (value) => {
+        if (typeof value === 'string') {
+          return value
+        }
+        if (value === null || value === undefined) {
+          return ''
+        }
+        // stream 对象存在循环引用，不能 JSON.stringify
+        if (typeof value === 'object' && typeof value.pipe === 'function') {
+          return ''
+        }
+        if (Buffer.isBuffer(value)) {
+          try {
+            return value.toString('utf8')
+          } catch (_) {
+            return ''
+          }
+        }
+        if (typeof value === 'object') {
+          try {
+            return JSON.stringify(value)
+          } catch (_) {
+            return ''
+          }
+        }
+        return String(value)
+      }
+
+      const msg = safeDataToString(data)
+      if (
+        msg.toLowerCase().includes('resource_exhausted') ||
+        msg.toLowerCase().includes('no capacity')
+      ) {
+        retriedAfterDelay = true
+        logger.warn('⏳ Antigravity 429 RESOURCE_EXHAUSTED, waiting 2s before retry', { model })
+        await new Promise((resolve) => setTimeout(resolve, 2000))
+        return await attemptRequest()
+      }
+    }
+    throw error
+  }
+}
+
+async function fetchAvailableModels({ accessToken, proxyConfig = null, timeoutMs = 30000 }) {
+  const proxyAgent = ProxyHelper.createProxyAgent(proxyConfig)
+  const endpoints = getAntigravityApiUrlCandidates()
+
+  let lastError = null
+  for (let index = 0; index < endpoints.length; index += 1) {
+    const baseUrl = endpoints[index]
+    const url = `${baseUrl}/v1internal:fetchAvailableModels`
+
+    const axiosConfig = {
+      url,
+      method: 'POST',
+      headers: getAntigravityHeaders(accessToken, baseUrl),
+      data: {},
+      timeout: timeoutMs
+    }
+
+    if (proxyAgent) {
+      axiosConfig.httpsAgent = proxyAgent
+      axiosConfig.proxy = false
+      if (index === 0) {
+        logger.info(
+          `🌐 Using proxy for Antigravity fetchAvailableModels: ${ProxyHelper.getProxyDescription(proxyConfig)}`
+        )
+      }
+    } else {
+      axiosConfig.httpsAgent = keepAliveAgent
+    }
+
+    try {
+      const response = await axios(axiosConfig)
+      return response.data
+    } catch (error) {
+      lastError = error
+      const status = error?.response?.status
+      const hasNext = index + 1 < endpoints.length
+      if (hasNext && (status === 429 || status === 404)) {
+        continue
+      }
+      throw error
+    }
+  }
+
+  throw lastError || new Error('Antigravity fetchAvailableModels failed')
+}
+
+async function countTokens({
+  accessToken,
+  proxyConfig = null,
+  contents,
+  model,
+  timeoutMs = 30000
+}) {
+  const upstreamModel = mapAntigravityUpstreamModel(model)
+
+  const proxyAgent = ProxyHelper.createProxyAgent(proxyConfig)
+  const endpoints = getAntigravityApiUrlCandidates()
+
+  let lastError = null
+  for (let index = 0; index < endpoints.length; index += 1) {
+    const baseUrl = endpoints[index]
+    const url = `${baseUrl}/v1internal:countTokens`
+    const axiosConfig = {
+      url,
+      method: 'POST',
+      headers: getAntigravityHeaders(accessToken, baseUrl),
+      data: {
+        request: {
+          model: `models/${upstreamModel}`,
+          contents
+        }
+      },
+      timeout: timeoutMs
+    }
+
+    if (proxyAgent) {
+      axiosConfig.httpsAgent = proxyAgent
+      axiosConfig.proxy = false
+      if (index === 0) {
+        logger.info(
+          `🌐 Using proxy for Antigravity countTokens: ${ProxyHelper.getProxyDescription(proxyConfig)}`
+        )
+      }
+    } else {
+      axiosConfig.httpsAgent = keepAliveAgent
+    }
+
+    try {
+      const response = await axios(axiosConfig)
+      return response.data
+    } catch (error) {
+      lastError = error
+      const status = error?.response?.status
+      const hasNext = index + 1 < endpoints.length
+      if (hasNext && (status === 429 || status === 404)) {
+        continue
+      }
+      throw error
+    }
+  }
+
+  throw lastError || new Error('Antigravity countTokens failed')
+}
+
+module.exports = {
+  getAntigravityApiUrl,
+  getAntigravityApiUrlCandidates,
+  getAntigravityHeaders,
+  buildAntigravityEnvelope,
+  request,
+  fetchAvailableModels,
+  countTokens
+}

src/services/antigravityRelayService.js

@@ -0,0 +1,170 @@
+const apiKeyService = require('./apiKeyService')
+const { convertMessagesToGemini, convertGeminiResponse } = require('./geminiRelayService')
+const { normalizeAntigravityModelInput } = require('../utils/antigravityModel')
+const antigravityClient = require('./antigravityClient')
+
+function buildRequestData({ messages, model, temperature, maxTokens, sessionId }) {
+  const requestedModel = normalizeAntigravityModelInput(model)
+  const { contents, systemInstruction } = convertMessagesToGemini(messages)
+
+  const requestData = {
+    model: requestedModel,
+    request: {
+      contents,
+      generationConfig: {
+        temperature,
+        maxOutputTokens: maxTokens,
+        candidateCount: 1,
+        topP: 0.95,
+        topK: 40
+      },
+      ...(sessionId ? { sessionId } : {})
+    }
+  }
+
+  if (systemInstruction) {
+    requestData.request.systemInstruction = { parts: [{ text: systemInstruction }] }
+  }
+
+  return requestData
+}
+
+async function* handleStreamResponse(response, model, apiKeyId, accountId) {
+  let buffer = ''
+  let totalUsage = {
+    promptTokenCount: 0,
+    candidatesTokenCount: 0,
+    totalTokenCount: 0
+  }
+  let usageRecorded = false
+
+  try {
+    for await (const chunk of response.data) {
+      buffer += chunk.toString()
+
+      const lines = buffer.split('\n')
+      buffer = lines.pop() || ''
+
+      for (const line of lines) {
+        if (!line.trim()) {
+          continue
+        }
+
+        let jsonData = line
+        if (line.startsWith('data: ')) {
+          jsonData = line.substring(6).trim()
+        }
+
+        if (!jsonData || jsonData === '[DONE]') {
+          continue
+        }
+
+        try {
+          const data = JSON.parse(jsonData)
+          const payload = data?.response || data
+
+          if (payload?.usageMetadata) {
+            totalUsage = payload.usageMetadata
+          }
+
+          const openaiChunk = convertGeminiResponse(payload, model, true)
+          if (openaiChunk) {
+            yield `data: ${JSON.stringify(openaiChunk)}\n\n`
+            const finishReason = openaiChunk.choices?.[0]?.finish_reason
+            if (finishReason === 'stop') {
+              yield 'data: [DONE]\n\n'
+
+              if (apiKeyId && totalUsage.totalTokenCount > 0) {
+                await apiKeyService.recordUsage(
+                  apiKeyId,
+                  totalUsage.promptTokenCount || 0,
+                  totalUsage.candidatesTokenCount || 0,
+                  0,
+                  0,
+                  model,
+                  accountId
+                )
+                usageRecorded = true
+              }
+              return
+            }
+          }
+        } catch (e) {
+          // ignore chunk parse errors
+        }
+      }
+    }
+  } finally {
+    if (!usageRecorded && apiKeyId && totalUsage.totalTokenCount > 0) {
+      await apiKeyService.recordUsage(
+        apiKeyId,
+        totalUsage.promptTokenCount || 0,
+        totalUsage.candidatesTokenCount || 0,
+        0,
+        0,
+        model,
+        accountId
+      )
+    }
+  }
+}
+
+async function sendAntigravityRequest({
+  messages,
+  model,
+  temperature = 0.7,
+  maxTokens = 4096,
+  stream = false,
+  accessToken,
+  proxy,
+  apiKeyId,
+  signal,
+  projectId,
+  accountId = null
+}) {
+  const requestedModel = normalizeAntigravityModelInput(model)
+
+  const requestData = buildRequestData({
+    messages,
+    model: requestedModel,
+    temperature,
+    maxTokens,
+    sessionId: apiKeyId
+  })
+
+  const { response } = await antigravityClient.request({
+    accessToken,
+    proxyConfig: proxy,
+    requestData,
+    projectId,
+    sessionId: apiKeyId,
+    stream,
+    signal,
+    params: { alt: 'sse' }
+  })
+
+  if (stream) {
+    return handleStreamResponse(response, requestedModel, apiKeyId, accountId)
+  }
+
+  const payload = response.data?.response || response.data
+  const openaiResponse = convertGeminiResponse(payload, requestedModel, false)
+
+  if (apiKeyId && openaiResponse?.usage) {
+    await apiKeyService.recordUsage(
+      apiKeyId,
+      openaiResponse.usage.prompt_tokens || 0,
+      openaiResponse.usage.completion_tokens || 0,
+      0,
+      0,
+      requestedModel,
+      accountId
+    )
+  }
+
+  return openaiResponse
+}
+
+module.exports = {
+  sendAntigravityRequest
+}

src/services/apiKeyService.js

@@ -37,6 +37,58 @@ const ACCOUNT_CATEGORY_MAP = {
   droid: 'droid'
 }
 
+/**
+ * 规范化权限数据，兼容旧格式（字符串）和新格式（数组）
+ * @param {string|array} permissions - 权限数据
+ * @returns {array} - 权限数组，空数组表示全部服务
+ */
+function normalizePermissions(permissions) {
+  if (!permissions) {
+    return [] // 空 = 全部服务
+  }
+  if (Array.isArray(permissions)) {
+    return permissions
+  }
+  // 尝试解析 JSON 字符串（新格式存储）
+  if (typeof permissions === 'string') {
+    if (permissions.startsWith('[')) {
+      try {
+        const parsed = JSON.parse(permissions)
+        if (Array.isArray(parsed)) {
+          return parsed
+        }
+      } catch (e) {
+        // 解析失败，继续处理为普通字符串
+      }
+    }
+    // 旧格式 'all' 转为空数组
+    if (permissions === 'all') {
+      return []
+    }
+    // 兼容逗号分隔格式（修复历史错误数据，如 "claude,openai"）
+    if (permissions.includes(',')) {
+      return permissions
+        .split(',')
+        .map((p) => p.trim())
+        .filter(Boolean)
+    }
+    // 旧单个字符串转为数组
+    return [permissions]
+  }
+  return []
+}
+
+/**
+ * 检查是否有访问特定服务的权限
+ * @param {string|array} permissions - 权限数据
+ * @param {string} service - 服务名称（claude/gemini/openai/droid）
+ * @returns {boolean} - 是否有权限
+ */
+function hasPermission(permissions, service) {
+  const perms = normalizePermissions(permissions)
+  return perms.length === 0 || perms.includes(service) // 空数组 = 全部服务
+}
+
 function normalizeAccountTypeKey(type) {
   if (!type) {
     return null
@@ -89,7 +141,7 @@ class ApiKeyService {
       azureOpenaiAccountId = null,
       bedrockAccountId = null, // 添加 Bedrock 账号ID支持
       droidAccountId = null,
-      permissions = 'all', // 可选值：'claude'、'gemini'、'openai'、'droid' 或 'all'
+      permissions = [], // 数组格式，空数组表示全部服务，如 ['claude', 'gemini']
       isActive = true,
       concurrencyLimit = 0,
       rateLimitWindow = null,
@@ -132,7 +184,7 @@ class ApiKeyService {
       azureOpenaiAccountId: azureOpenaiAccountId || '',
       bedrockAccountId: bedrockAccountId || '', // 添加 Bedrock 账号ID
       droidAccountId: droidAccountId || '',
-      permissions: permissions || 'all',
+      permissions: JSON.stringify(normalizePermissions(permissions)),
       enableModelRestriction: String(enableModelRestriction),
       restrictedModels: JSON.stringify(restrictedModels || []),
       enableClientRestriction: String(enableClientRestriction || false),
@@ -186,7 +238,7 @@ class ApiKeyService {
       azureOpenaiAccountId: keyData.azureOpenaiAccountId,
       bedrockAccountId: keyData.bedrockAccountId, // 添加 Bedrock 账号ID
       droidAccountId: keyData.droidAccountId,
-      permissions: keyData.permissions,
+      permissions: normalizePermissions(keyData.permissions),
       enableModelRestriction: keyData.enableModelRestriction === 'true',
       restrictedModels: JSON.parse(keyData.restrictedModels),
       enableClientRestriction: keyData.enableClientRestriction === 'true',
@@ -338,7 +390,7 @@ class ApiKeyService {
           azureOpenaiAccountId: keyData.azureOpenaiAccountId,
           bedrockAccountId: keyData.bedrockAccountId, // 添加 Bedrock 账号ID
           droidAccountId: keyData.droidAccountId,
-          permissions: keyData.permissions || 'all',
+          permissions: normalizePermissions(keyData.permissions),
           tokenLimit: parseInt(keyData.tokenLimit),
           concurrencyLimit: parseInt(keyData.concurrencyLimit || 0),
           rateLimitWindow: parseInt(keyData.rateLimitWindow || 0),
@@ -467,7 +519,7 @@ class ApiKeyService {
           azureOpenaiAccountId: keyData.azureOpenaiAccountId,
           bedrockAccountId: keyData.bedrockAccountId,
           droidAccountId: keyData.droidAccountId,
-          permissions: keyData.permissions || 'all',
+          permissions: normalizePermissions(keyData.permissions),
           tokenLimit: parseInt(keyData.tokenLimit),
           concurrencyLimit: parseInt(keyData.concurrencyLimit || 0),
           rateLimitWindow: parseInt(keyData.rateLimitWindow || 0),
@@ -525,7 +577,7 @@ class ApiKeyService {
         key.isActive = key.isActive === 'true'
         key.enableModelRestriction = key.enableModelRestriction === 'true'
         key.enableClientRestriction = key.enableClientRestriction === 'true'
-        key.permissions = key.permissions || 'all' // 兼容旧数据
+        key.permissions = normalizePermissions(key.permissions)
         key.dailyCostLimit = parseFloat(key.dailyCostLimit || 0)
         key.totalCostLimit = parseFloat(key.totalCostLimit || 0)
         key.weeklyOpusCostLimit = parseFloat(key.weeklyOpusCostLimit || 0)
@@ -708,6 +760,9 @@ class ApiKeyService {
           if (field === 'restrictedModels' || field === 'allowedClients' || field === 'tags') {
             // 特殊处理数组字段
             updatedData[field] = JSON.stringify(value || [])
+          } else if (field === 'permissions') {
+            // 权限字段：规范化后JSON序列化，与createApiKey保持一致
+            updatedData[field] = JSON.stringify(normalizePermissions(value))
           } else if (
             field === 'enableModelRestriction' ||
             field === 'enableClientRestriction' ||
@@ -1568,7 +1623,7 @@ class ApiKeyService {
         userId: keyData.userId,
         userUsername: keyData.userUsername,
         createdBy: keyData.createdBy,
-        permissions: keyData.permissions,
+        permissions: normalizePermissions(keyData.permissions),
         dailyCostLimit: parseFloat(keyData.dailyCostLimit || 0),
         totalCostLimit: parseFloat(keyData.totalCostLimit || 0),
         // 所有平台账户绑定字段
@@ -1820,4 +1875,8 @@ const apiKeyService = new ApiKeyService()
 // 为了方便其他服务调用，导出 recordUsage 方法
 apiKeyService.recordUsageMetrics = apiKeyService.recordUsage.bind(apiKeyService)
 
+// 导出权限辅助函数供路由使用
+apiKeyService.hasPermission = hasPermission
+apiKeyService.normalizePermissions = normalizePermissions
+
 module.exports = apiKeyService

src/services/balanceProviders/baseBalanceProvider.js

@@ -0,0 +1,133 @@
+const axios = require('axios')
+const logger = require('../../utils/logger')
+const ProxyHelper = require('../../utils/proxyHelper')
+
+/**
+ * Provider 抽象基类
+ * 各平台 Provider 需继承并实现 queryBalance(account)
+ */
+class BaseBalanceProvider {
+  constructor(platform) {
+    this.platform = platform
+    this.logger = logger
+  }
+
+  /**
+   * 查询余额（抽象方法）
+   * @param {object} account - 账户对象
+   * @returns {Promise<object>}
+   * 形如：
+   * {
+   *   balance: number|null,
+   *   currency?: string,
+   *   quota?: { daily, used, remaining, resetAt, percentage, unlimited? },
+   *   queryMethod?: 'api'|'field'|'local',
+   *   rawData?: any
+   * }
+   */
+  async queryBalance(_account) {
+    throw new Error('queryBalance 方法必须由子类实现')
+  }
+
+  /**
+   * 通用 HTTP 请求方法（支持代理）
+   * @param {string} url
+   * @param {object} options
+   * @param {object} account
+   */
+  async makeRequest(url, options = {}, account = {}) {
+    const config = {
+      url,
+      method: options.method || 'GET',
+      headers: options.headers || {},
+      timeout: options.timeout || 15000,
+      data: options.data,
+      params: options.params,
+      responseType: options.responseType
+    }
+
+    const proxyConfig = account.proxyConfig || account.proxy
+    if (proxyConfig) {
+      const agent = ProxyHelper.createProxyAgent(proxyConfig)
+      if (agent) {
+        config.httpAgent = agent
+        config.httpsAgent = agent
+        config.proxy = false
+      }
+    }
+
+    try {
+      const response = await axios(config)
+      return {
+        success: true,
+        data: response.data,
+        status: response.status,
+        headers: response.headers
+      }
+    } catch (error) {
+      const status = error.response?.status
+      const message = error.response?.data?.message || error.message || '请求失败'
+      this.logger.debug(`余额 Provider HTTP 请求失败: ${url} (${this.platform})`, {
+        status,
+        message
+      })
+      return { success: false, status, error: message }
+    }
+  }
+
+  /**
+   * 从账户字段读取 dailyQuota / dailyUsage（通用降级方案）
+   * 注意：部分平台 dailyUsage 字段可能不是实时值，最终以 AccountBalanceService 的本地统计为准
+   */
+  readQuotaFromFields(account) {
+    const dailyQuota = Number(account?.dailyQuota || 0)
+    const dailyUsage = Number(account?.dailyUsage || 0)
+
+    // 无限制
+    if (!Number.isFinite(dailyQuota) || dailyQuota <= 0) {
+      return {
+        balance: null,
+        currency: 'USD',
+        quota: {
+          daily: Infinity,
+          used: Number.isFinite(dailyUsage) ? dailyUsage : 0,
+          remaining: Infinity,
+          percentage: 0,
+          unlimited: true
+        },
+        queryMethod: 'field'
+      }
+    }
+
+    const used = Number.isFinite(dailyUsage) ? dailyUsage : 0
+    const remaining = Math.max(0, dailyQuota - used)
+    const percentage = dailyQuota > 0 ? (used / dailyQuota) * 100 : 0
+
+    return {
+      balance: remaining,
+      currency: 'USD',
+      quota: {
+        daily: dailyQuota,
+        used,
+        remaining,
+        percentage: Math.round(percentage * 100) / 100
+      },
+      queryMethod: 'field'
+    }
+  }
+
+  parseCurrency(data) {
+    return data?.currency || data?.Currency || 'USD'
+  }
+
+  async safeExecute(fn, fallbackValue = null) {
+    try {
+      return await fn()
+    } catch (error) {
+      this.logger.error(`余额 Provider 执行失败: ${this.platform}`, error)
+      return fallbackValue
+    }
+  }
+}
+
+module.exports = BaseBalanceProvider

src/services/balanceProviders/claudeBalanceProvider.js

@@ -0,0 +1,30 @@
+const BaseBalanceProvider = require('./baseBalanceProvider')
+const claudeAccountService = require('../claudeAccountService')
+
+class ClaudeBalanceProvider extends BaseBalanceProvider {
+  constructor() {
+    super('claude')
+  }
+
+  /**
+   * Claude（OAuth）：优先尝试获取 OAuth usage（用于配额/使用信息），不强行提供余额金额
+   */
+  async queryBalance(account) {
+    this.logger.debug(`查询 Claude 余额（OAuth usage）: ${account?.id}`)
+
+    // 仅 OAuth 账户可用；失败时降级
+    const usageData = await claudeAccountService.fetchOAuthUsage(account.id).catch(() => null)
+    if (!usageData) {
+      return { balance: null, currency: 'USD', queryMethod: 'local' }
+    }
+
+    return {
+      balance: null,
+      currency: 'USD',
+      queryMethod: 'api',
+      rawData: usageData
+    }
+  }
+}
+
+module.exports = ClaudeBalanceProvider

src/services/balanceProviders/claudeConsoleBalanceProvider.js

@@ -0,0 +1,14 @@
+const BaseBalanceProvider = require('./baseBalanceProvider')
+
+class ClaudeConsoleBalanceProvider extends BaseBalanceProvider {
+  constructor() {
+    super('claude-console')
+  }
+
+  async queryBalance(account) {
+    this.logger.debug(`查询 Claude Console 余额（字段）: ${account?.id}`)
+    return this.readQuotaFromFields(account)
+  }
+}
+
+module.exports = ClaudeConsoleBalanceProvider

src/services/balanceProviders/geminiBalanceProvider.js

@@ -0,0 +1,250 @@
+const BaseBalanceProvider = require('./baseBalanceProvider')
+const antigravityClient = require('../antigravityClient')
+const geminiAccountService = require('../geminiAccountService')
+
+const OAUTH_PROVIDER_ANTIGRAVITY = 'antigravity'
+
+function clamp01(value) {
+  if (typeof value !== 'number' || !Number.isFinite(value)) {
+    return null
+  }
+  if (value < 0) {
+    return 0
+  }
+  if (value > 1) {
+    return 1
+  }
+  return value
+}
+
+function round2(value) {
+  if (typeof value !== 'number' || !Number.isFinite(value)) {
+    return null
+  }
+  return Math.round(value * 100) / 100
+}
+
+function normalizeQuotaCategory(displayName, modelId) {
+  const name = String(displayName || '')
+  const id = String(modelId || '')
+
+  if (name.includes('Gemini') && name.includes('Pro')) {
+    return 'Gemini Pro'
+  }
+  if (name.includes('Gemini') && name.includes('Flash')) {
+    return 'Gemini Flash'
+  }
+  if (name.includes('Gemini') && name.toLowerCase().includes('image')) {
+    return 'Gemini Image'
+  }
+
+  if (name.includes('Claude') || name.includes('GPT-OSS')) {
+    return 'Claude'
+  }
+
+  if (id.startsWith('gemini-3-pro-') || id.startsWith('gemini-2.5-pro')) {
+    return 'Gemini Pro'
+  }
+  if (id.startsWith('gemini-3-flash') || id.startsWith('gemini-2.5-flash')) {
+    return 'Gemini Flash'
+  }
+  if (id.includes('image')) {
+    return 'Gemini Image'
+  }
+  if (id.includes('claude') || id.includes('gpt-oss')) {
+    return 'Claude'
+  }
+
+  return name || id || 'Unknown'
+}
+
+function buildAntigravityQuota(modelsResponse) {
+  const models = modelsResponse && typeof modelsResponse === 'object' ? modelsResponse.models : null
+
+  if (!models || typeof models !== 'object') {
+    return null
+  }
+
+  const parseRemainingFraction = (quotaInfo) => {
+    if (!quotaInfo || typeof quotaInfo !== 'object') {
+      return null
+    }
+
+    const raw =
+      quotaInfo.remainingFraction ??
+      quotaInfo.remaining_fraction ??
+      quotaInfo.remaining ??
+      undefined
+
+    const num = typeof raw === 'number' ? raw : typeof raw === 'string' ? Number(raw) : NaN
+    if (!Number.isFinite(num)) {
+      return null
+    }
+
+    return clamp01(num)
+  }
+
+  const allowedCategories = new Set(['Gemini Pro', 'Claude', 'Gemini Flash', 'Gemini Image'])
+  const fixedOrder = ['Gemini Pro', 'Claude', 'Gemini Flash', 'Gemini Image']
+
+  const categoryMap = new Map()
+
+  for (const [modelId, modelDataRaw] of Object.entries(models)) {
+    if (!modelDataRaw || typeof modelDataRaw !== 'object') {
+      continue
+    }
+
+    const displayName = modelDataRaw.displayName || modelDataRaw.display_name || modelId
+    const quotaInfo = modelDataRaw.quotaInfo || modelDataRaw.quota_info || null
+
+    const remainingFraction = parseRemainingFraction(quotaInfo)
+    if (remainingFraction === null) {
+      continue
+    }
+
+    const remainingPercent = round2(remainingFraction * 100)
+    const usedPercent = round2(100 - remainingPercent)
+    const resetAt = quotaInfo?.resetTime || quotaInfo?.reset_time || null
+
+    const category = normalizeQuotaCategory(displayName, modelId)
+    if (!allowedCategories.has(category)) {
+      continue
+    }
+    const entry = {
+      category,
+      modelId,
+      displayName: String(displayName || modelId || category),
+      remainingPercent,
+      usedPercent,
+      resetAt: typeof resetAt === 'string' && resetAt.trim() ? resetAt : null
+    }
+
+    const existing = categoryMap.get(category)
+    if (!existing || entry.remainingPercent < existing.remainingPercent) {
+      categoryMap.set(category, entry)
+    }
+  }
+
+  const buckets = fixedOrder.map((category) => {
+    const existing = categoryMap.get(category) || null
+    if (existing) {
+      return existing
+    }
+    return {
+      category,
+      modelId: '',
+      displayName: category,
+      remainingPercent: null,
+      usedPercent: null,
+      resetAt: null
+    }
+  })
+
+  if (buckets.length === 0) {
+    return null
+  }
+
+  const critical = buckets
+    .filter((item) => item.remainingPercent !== null)
+    .reduce((min, item) => {
+      if (!min) {
+        return item
+      }
+      return (item.remainingPercent ?? 0) < (min.remainingPercent ?? 0) ? item : min
+    }, null)
+
+  if (!critical) {
+    return null
+  }
+
+  return {
+    balance: null,
+    currency: 'USD',
+    quota: {
+      type: 'antigravity',
+      total: 100,
+      used: critical.usedPercent,
+      remaining: critical.remainingPercent,
+      percentage: critical.usedPercent,
+      resetAt: critical.resetAt,
+      buckets: buckets.map((item) => ({
+        category: item.category,
+        remaining: item.remainingPercent,
+        used: item.usedPercent,
+        percentage: item.usedPercent,
+        resetAt: item.resetAt
+      }))
+    },
+    queryMethod: 'api',
+    rawData: {
+      modelsCount: Object.keys(models).length,
+      bucketCount: buckets.length
+    }
+  }
+}
+
+class GeminiBalanceProvider extends BaseBalanceProvider {
+  constructor() {
+    super('gemini')
+  }
+
+  async queryBalance(account) {
+    const oauthProvider = account?.oauthProvider
+    if (oauthProvider !== OAUTH_PROVIDER_ANTIGRAVITY) {
+      if (account && Object.prototype.hasOwnProperty.call(account, 'dailyQuota')) {
+        return this.readQuotaFromFields(account)
+      }
+      return { balance: null, currency: 'USD', queryMethod: 'local' }
+    }
+
+    const accessToken = String(account?.accessToken || '').trim()
+    const refreshToken = String(account?.refreshToken || '').trim()
+    const proxyConfig = account?.proxyConfig || account?.proxy || null
+
+    if (!accessToken) {
+      throw new Error('Antigravity 账户缺少 accessToken')
+    }
+
+    const fetch = async (token) =>
+      await antigravityClient.fetchAvailableModels({
+        accessToken: token,
+        proxyConfig
+      })
+
+    let data
+    try {
+      data = await fetch(accessToken)
+    } catch (error) {
+      const status = error?.response?.status
+      if ((status === 401 || status === 403) && refreshToken) {
+        const refreshed = await geminiAccountService.refreshAccessToken(
+          refreshToken,
+          proxyConfig,
+          OAUTH_PROVIDER_ANTIGRAVITY
+        )
+        const nextToken = String(refreshed?.access_token || '').trim()
+        if (!nextToken) {
+          throw error
+        }
+        data = await fetch(nextToken)
+      } else {
+        throw error
+      }
+    }
+
+    const mapped = buildAntigravityQuota(data)
+    if (!mapped) {
+      return {
+        balance: null,
+        currency: 'USD',
+        quota: null,
+        queryMethod: 'api',
+        rawData: data || null
+      }
+    }
+
+    return mapped
+  }
+}
+
+module.exports = GeminiBalanceProvider

src/services/balanceProviders/genericBalanceProvider.js

@@ -0,0 +1,23 @@
+const BaseBalanceProvider = require('./baseBalanceProvider')
+
+class GenericBalanceProvider extends BaseBalanceProvider {
+  constructor(platform) {
+    super(platform)
+  }
+
+  async queryBalance(account) {
+    this.logger.debug(`${this.platform} 暂无专用余额 API，实现降级策略`)
+
+    if (account && Object.prototype.hasOwnProperty.call(account, 'dailyQuota')) {
+      return this.readQuotaFromFields(account)
+    }
+
+    return {
+      balance: null,
+      currency: 'USD',
+      queryMethod: 'local'
+    }
+  }
+}
+
+module.exports = GenericBalanceProvider

src/services/balanceProviders/index.js

@@ -0,0 +1,25 @@
+const ClaudeBalanceProvider = require('./claudeBalanceProvider')
+const ClaudeConsoleBalanceProvider = require('./claudeConsoleBalanceProvider')
+const OpenAIResponsesBalanceProvider = require('./openaiResponsesBalanceProvider')
+const GenericBalanceProvider = require('./genericBalanceProvider')
+const GeminiBalanceProvider = require('./geminiBalanceProvider')
+
+function registerAllProviders(balanceService) {
+  // Claude
+  balanceService.registerProvider('claude', new ClaudeBalanceProvider())
+  balanceService.registerProvider('claude-console', new ClaudeConsoleBalanceProvider())
+
+  // OpenAI / Codex
+  balanceService.registerProvider('openai-responses', new OpenAIResponsesBalanceProvider())
+  balanceService.registerProvider('openai', new GenericBalanceProvider('openai'))
+  balanceService.registerProvider('azure_openai', new GenericBalanceProvider('azure_openai'))
+
+  // 其他平台（降级）
+  balanceService.registerProvider('gemini', new GeminiBalanceProvider())
+  balanceService.registerProvider('gemini-api', new GenericBalanceProvider('gemini-api'))
+  balanceService.registerProvider('bedrock', new GenericBalanceProvider('bedrock'))
+  balanceService.registerProvider('droid', new GenericBalanceProvider('droid'))
+  balanceService.registerProvider('ccr', new GenericBalanceProvider('ccr'))
+}
+
+module.exports = { registerAllProviders }

src/services/balanceProviders/openaiResponsesBalanceProvider.js

@@ -0,0 +1,54 @@
+const BaseBalanceProvider = require('./baseBalanceProvider')
+
+class OpenAIResponsesBalanceProvider extends BaseBalanceProvider {
+  constructor() {
+    super('openai-responses')
+  }
+
+  /**
+   * OpenAI-Responses：
+   * - 优先使用 dailyQuota 字段（如果配置了额度）
+   * - 可选：尝试调用兼容 API（不同服务商实现不一，失败自动降级）
+   */
+  async queryBalance(account) {
+    this.logger.debug(`查询 OpenAI Responses 余额: ${account?.id}`)
+
+    // 配置了额度时直接返回（字段法）
+    if (account?.dailyQuota && Number(account.dailyQuota) > 0) {
+      return this.readQuotaFromFields(account)
+    }
+
+    // 尝试调用 usage 接口（兼容性不保证）
+    if (account?.apiKey && account?.baseApi) {
+      const baseApi = String(account.baseApi).replace(/\/$/, '')
+      const response = await this.makeRequest(
+        `${baseApi}/v1/usage`,
+        {
+          method: 'GET',
+          headers: {
+            Authorization: `Bearer ${account.apiKey}`,
+            'Content-Type': 'application/json'
+          }
+        },
+        account
+      )
+
+      if (response.success) {
+        return {
+          balance: null,
+          currency: this.parseCurrency(response.data),
+          queryMethod: 'api',
+          rawData: response.data
+        }
+      }
+    }
+
+    return {
+      balance: null,
+      currency: 'USD',
+      queryMethod: 'local'
+    }
+  }
+}
+
+module.exports = OpenAIResponsesBalanceProvider

src/services/balanceScriptService.js

@@ -0,0 +1,210 @@
+const vm = require('vm')
+const axios = require('axios')
+const { isBalanceScriptEnabled } = require('../utils/featureFlags')
+
+/**
+ * SSRF防护：检查URL是否访问内网或敏感地址
+ * @param {string} url - 要检查的URL
+ * @returns {boolean} - true表示URL安全
+ */
+function isUrlSafe(url) {
+  try {
+    const parsed = new URL(url)
+    const hostname = parsed.hostname.toLowerCase()
+
+    // 禁止的协议
+    if (!['http:', 'https:'].includes(parsed.protocol)) {
+      return false
+    }
+
+    // 禁止访问localhost和私有IP
+    const privatePatterns = [
+      /^localhost$/i,
+      /^127\./,
+      /^10\./,
+      /^172\.(1[6-9]|2[0-9]|3[0-1])\./,
+      /^192\.168\./,
+      /^169\.254\./, // AWS metadata
+      /^0\./, // 0.0.0.0
+      /^::1$/,
+      /^fc00:/i,
+      /^fe80:/i,
+      /\.local$/i,
+      /\.internal$/i,
+      /\.localhost$/i
+    ]
+
+    for (const pattern of privatePatterns) {
+      if (pattern.test(hostname)) {
+        return false
+      }
+    }
+
+    return true
+  } catch {
+    return false
+  }
+}
+
+/**
+ * 可配置脚本余额查询执行器
+ * - 脚本格式：({ request: {...}, extractor: function(response){...} })
+ * - 模板变量：{{baseUrl}}, {{apiKey}}, {{token}}, {{accountId}}, {{platform}}, {{extra}}
+ */
+class BalanceScriptService {
+  /**
+   * 执行脚本：返回标准余额结构 + 原始响应
+   * @param {object} options
+   *  - scriptBody: string
+   *  - variables: Record<string,string>
+   *  - timeoutSeconds: number
+   */
+  async execute(options = {}) {
+    if (!isBalanceScriptEnabled()) {
+      const error = new Error('余额脚本功能已禁用（可通过 BALANCE_SCRIPT_ENABLED=true 启用）')
+      error.code = 'BALANCE_SCRIPT_DISABLED'
+      throw error
+    }
+
+    const scriptBody = options.scriptBody?.trim()
+    if (!scriptBody) {
+      throw new Error('脚本内容为空')
+    }
+
+    const timeoutMs = Math.max(1, (options.timeoutSeconds || 10) * 1000)
+    const sandbox = {
+      console,
+      Math,
+      Date
+    }
+
+    let scriptResult
+    try {
+      const wrapped = scriptBody.startsWith('(') ? scriptBody : `(${scriptBody})`
+      const script = new vm.Script(wrapped)
+      scriptResult = script.runInNewContext(sandbox, { timeout: timeoutMs })
+    } catch (error) {
+      throw new Error(`脚本解析失败: ${error.message}`)
+    }
+
+    if (!scriptResult || typeof scriptResult !== 'object') {
+      throw new Error('脚本返回格式无效（需返回 { request, extractor }）')
+    }
+
+    const variables = options.variables || {}
+    const request = this.applyTemplates(scriptResult.request || {}, variables)
+    const { extractor } = scriptResult
+
+    if (!request?.url || typeof request.url !== 'string') {
+      throw new Error('脚本 request.url 不能为空')
+    }
+
+    // SSRF防护：验证URL安全性
+    if (!isUrlSafe(request.url)) {
+      throw new Error('脚本 request.url 不安全：禁止访问内网地址、localhost或使用非HTTP(S)协议')
+    }
+
+    if (typeof extractor !== 'function') {
+      throw new Error('脚本 extractor 必须是函数')
+    }
+
+    const axiosConfig = {
+      url: request.url,
+      method: (request.method || 'GET').toUpperCase(),
+      headers: request.headers || {},
+      timeout: timeoutMs
+    }
+
+    if (request.params) {
+      axiosConfig.params = request.params
+    }
+    if (request.body || request.data) {
+      axiosConfig.data = request.body || request.data
+    }
+
+    let httpResponse
+    try {
+      httpResponse = await axios(axiosConfig)
+    } catch (error) {
+      const { response } = error || {}
+      const { status, data } = response || {}
+      throw new Error(
+        `请求失败: ${status || ''} ${error.message}${data ? ` | ${JSON.stringify(data)}` : ''}`
+      )
+    }
+
+    const responseData = httpResponse?.data
+
+    let extracted = {}
+    try {
+      extracted = extractor(responseData) || {}
+    } catch (error) {
+      throw new Error(`extractor 执行失败: ${error.message}`)
+    }
+
+    const mapped = this.mapExtractorResult(extracted, responseData)
+    return {
+      mapped,
+      extracted,
+      response: {
+        status: httpResponse?.status,
+        headers: httpResponse?.headers,
+        data: responseData
+      }
+    }
+  }
+
+  applyTemplates(value, variables) {
+    if (typeof value === 'string') {
+      return value.replace(/{{(\w+)}}/g, (_, key) => {
+        const trimmed = key.trim()
+        return variables[trimmed] !== undefined ? String(variables[trimmed]) : ''
+      })
+    }
+    if (Array.isArray(value)) {
+      return value.map((item) => this.applyTemplates(item, variables))
+    }
+    if (value && typeof value === 'object') {
+      const result = {}
+      Object.keys(value).forEach((k) => {
+        result[k] = this.applyTemplates(value[k], variables)
+      })
+      return result
+    }
+    return value
+  }
+
+  mapExtractorResult(result = {}, responseData) {
+    const isValid = result.isValid !== false
+    const remaining = Number(result.remaining)
+    const total = Number(result.total)
+    const used = Number(result.used)
+    const currency = result.unit || 'USD'
+
+    const quota =
+      Number.isFinite(total) || Number.isFinite(used)
+        ? {
+            total: Number.isFinite(total) ? total : null,
+            used: Number.isFinite(used) ? used : null,
+            remaining: Number.isFinite(remaining) ? remaining : null,
+            percentage:
+              Number.isFinite(total) && total > 0 && Number.isFinite(used)
+                ? (used / total) * 100
+                : null
+          }
+        : null
+
+    return {
+      status: isValid ? 'success' : 'error',
+      errorMessage: isValid ? '' : result.invalidMessage || '套餐无效',
+      balance: Number.isFinite(remaining) ? remaining : null,
+      currency,
+      quota,
+      planName: result.planName || null,
+      extra: result.extra || null,
+      rawData: responseData || result.raw
+    }
+  }
+}
+
+module.exports = new BalanceScriptService()

src/services/bedrockAccountService.js

@@ -35,12 +35,13 @@ class BedrockAccountService {
       description = '',
       region = process.env.AWS_REGION || 'us-east-1',
       awsCredentials = null, // { accessKeyId, secretAccessKey, sessionToken }
+      bearerToken = null, // AWS Bearer Token for Bedrock API Keys
       defaultModel = 'us.anthropic.claude-sonnet-4-20250514-v1:0',
       isActive = true,
       accountType = 'shared', // 'dedicated' or 'shared'
       priority = 50, // 调度优先级 (1-100，数字越小优先级越高)
       schedulable = true, // 是否可被调度
-      credentialType = 'default' // 'default', 'access_key', 'bearer_token'
+      credentialType = 'access_key' // 'access_key', 'bearer_token'（默认为 access_key）
     } = options
 
     const accountId = uuidv4()
@@ -71,6 +72,11 @@ class BedrockAccountService {
       accountData.awsCredentials = this._encryptAwsCredentials(awsCredentials)
     }
 
+    // 加密存储 Bearer Token
+    if (bearerToken) {
+      accountData.bearerToken = this._encryptAwsCredentials({ token: bearerToken })
+    }
+
     const client = redis.getClientSafe()
     await client.set(`bedrock_account:${accountId}`, JSON.stringify(accountData))
 
@@ -106,9 +112,85 @@ class BedrockAccountService {
 
       const account = JSON.parse(accountData)
 
-      // 解密AWS凭证用于内部使用
-      if (account.awsCredentials) {
-        account.awsCredentials = this._decryptAwsCredentials(account.awsCredentials)
+      // 根据凭证类型解密对应的凭证
+      // 增强逻辑：优先按照 credentialType 解密，如果字段不存在则尝试解密实际存在的字段（兜底）
+      try {
+        let accessKeyDecrypted = false
+        let bearerTokenDecrypted = false
+
+        // 第一步：按照 credentialType 尝试解密对应的凭证
+        if (account.credentialType === 'access_key' && account.awsCredentials) {
+          // Access Key 模式：解密 AWS 凭证
+          account.awsCredentials = this._decryptAwsCredentials(account.awsCredentials)
+          accessKeyDecrypted = true
+          logger.debug(
+            `🔓 解密 Access Key 成功 - ID: ${accountId}, 类型: ${account.credentialType}`
+          )
+        } else if (account.credentialType === 'bearer_token' && account.bearerToken) {
+          // Bearer Token 模式：解密 Bearer Token
+          const decrypted = this._decryptAwsCredentials(account.bearerToken)
+          account.bearerToken = decrypted.token
+          bearerTokenDecrypted = true
+          logger.debug(
+            `🔓 解密 Bearer Token 成功 - ID: ${accountId}, 类型: ${account.credentialType}`
+          )
+        } else if (!account.credentialType || account.credentialType === 'default') {
+          // 向后兼容：旧版本账号可能没有 credentialType 字段，尝试解密所有存在的凭证
+          if (account.awsCredentials) {
+            account.awsCredentials = this._decryptAwsCredentials(account.awsCredentials)
+            accessKeyDecrypted = true
+          }
+          if (account.bearerToken) {
+            const decrypted = this._decryptAwsCredentials(account.bearerToken)
+            account.bearerToken = decrypted.token
+            bearerTokenDecrypted = true
+          }
+          logger.debug(
+            `🔓 兼容模式解密 - ID: ${accountId}, Access Key: ${accessKeyDecrypted}, Bearer Token: ${bearerTokenDecrypted}`
+          )
+        }
+
+        // 第二步：兜底逻辑 - 如果按照 credentialType 没有解密到任何凭证，尝试解密实际存在的字段
+        if (!accessKeyDecrypted && !bearerTokenDecrypted) {
+          logger.warn(
+            `⚠️ credentialType="${account.credentialType}" 与实际字段不匹配，尝试兜底解密 - ID: ${accountId}`
+          )
+          if (account.awsCredentials) {
+            account.awsCredentials = this._decryptAwsCredentials(account.awsCredentials)
+            accessKeyDecrypted = true
+            logger.warn(
+              `🔓 兜底解密 Access Key 成功 - ID: ${accountId}, credentialType 应为 'access_key'`
+            )
+          }
+          if (account.bearerToken) {
+            const decrypted = this._decryptAwsCredentials(account.bearerToken)
+            account.bearerToken = decrypted.token
+            bearerTokenDecrypted = true
+            logger.warn(
+              `🔓 兜底解密 Bearer Token 成功 - ID: ${accountId}, credentialType 应为 'bearer_token'`
+            )
+          }
+        }
+
+        // 验证至少解密了一种凭证
+        if (!accessKeyDecrypted && !bearerTokenDecrypted) {
+          logger.error(
+            `❌ 未找到任何凭证可解密 - ID: ${accountId}, credentialType: ${account.credentialType}, hasAwsCredentials: ${!!account.awsCredentials}, hasBearerToken: ${!!account.bearerToken}`
+          )
+          return {
+            success: false,
+            error: 'No valid credentials found in account data'
+          }
+        }
+      } catch (decryptError) {
+        logger.error(
+          `❌ 解密Bedrock凭证失败 - ID: ${accountId}, 类型: ${account.credentialType}`,
+          decryptError
+        )
+        return {
+          success: false,
+          error: `Credentials decryption failed: ${decryptError.message}`
+        }
       }
 
       logger.debug(`🔍 获取Bedrock账户 - ID: ${accountId}, 名称: ${account.name}`)
@@ -155,7 +237,11 @@ class BedrockAccountService {
             updatedAt: account.updatedAt,
             type: 'bedrock',
             platform: 'bedrock',
-            hasCredentials: !!account.awsCredentials
+            // 根据凭证类型判断是否有凭证
+            hasCredentials:
+              account.credentialType === 'bearer_token'
+                ? !!account.bearerToken
+                : !!account.awsCredentials
           })
         }
       }
@@ -235,6 +321,15 @@ class BedrockAccountService {
         logger.info(`🔐 重新加密Bedrock账户凭证 - ID: ${accountId}`)
       }
 
+      // 更新 Bearer Token
+      if (updates.bearerToken !== undefined) {
+        if (updates.bearerToken) {
+          account.bearerToken = this._encryptAwsCredentials({ token: updates.bearerToken })
+        } else {
+          delete account.bearerToken
+        }
+      }
+
       // ✅ 直接保存 subscriptionExpiresAt（如果提供）
       // Bedrock 没有 token 刷新逻辑，不会覆盖此字段
       if (updates.subscriptionExpiresAt !== undefined) {
@@ -345,13 +440,45 @@ class BedrockAccountService {
 
       const account = accountResult.data
 
-      logger.info(`🧪 测试Bedrock账户连接 - ID: ${accountId}, 名称: ${account.name}`)
+      logger.info(
+        `🧪 测试Bedrock账户连接 - ID: ${accountId}, 名称: ${account.name}, 凭证类型: ${account.credentialType}`
+      )
 
-      // 尝试获取模型列表来测试连接
+      // 验证凭证是否已解密
+      const hasValidCredentials =
+        (account.credentialType === 'access_key' && account.awsCredentials) ||
+        (account.credentialType === 'bearer_token' && account.bearerToken) ||
+        (!account.credentialType && (account.awsCredentials || account.bearerToken))
+
+      if (!hasValidCredentials) {
+        logger.error(
+          `❌ 测试失败：账户没有有效凭证 - ID: ${accountId}, credentialType: ${account.credentialType}`
+        )
+        return {
+          success: false,
+          error: 'No valid credentials found after decryption'
+        }
+      }
+
+      // 尝试创建 Bedrock 客户端来验证凭证格式
+      try {
+        bedrockRelayService._getBedrockClient(account.region, account)
+        logger.debug(`✅ Bedrock客户端创建成功 - ID: ${accountId}`)
+      } catch (clientError) {
+        logger.error(`❌ 创建Bedrock客户端失败 - ID: ${accountId}`, clientError)
+        return {
+          success: false,
+          error: `Failed to create Bedrock client: ${clientError.message}`
+        }
+      }
+
+      // 获取可用模型列表（硬编码，但至少验证了凭证格式正确）
       const models = await bedrockRelayService.getAvailableModels(account)
 
       if (models && models.length > 0) {
-        logger.info(`✅ Bedrock账户测试成功 - ID: ${accountId}, 发现 ${models.length} 个模型`)
+        logger.info(
+          `✅ Bedrock账户测试成功 - ID: ${accountId}, 发现 ${models.length} 个模型, 凭证类型: ${account.credentialType}`
+        )
         return {
           success: true,
           data: {
@@ -376,6 +503,135 @@ class BedrockAccountService {
     }
   }
 
+  /**
+   * 🧪 测试 Bedrock 账户连接（SSE 流式返回，供前端测试页面使用）
+   * @param {string} accountId - 账户ID
+   * @param {Object} res - Express response 对象
+   * @param {string} model - 测试使用的模型
+   */
+  async testAccountConnection(accountId, res, model = null) {
+    const { InvokeModelWithResponseStreamCommand } = require('@aws-sdk/client-bedrock-runtime')
+
+    try {
+      // 获取账户信息
+      const accountResult = await this.getAccount(accountId)
+      if (!accountResult.success) {
+        throw new Error(accountResult.error || 'Account not found')
+      }
+
+      const account = accountResult.data
+
+      // 根据账户类型选择合适的测试模型
+      if (!model) {
+        // Access Key 模式使用 Haiku（更快更便宜）
+        model = account.defaultModel || 'us.anthropic.claude-3-5-haiku-20241022-v1:0'
+      }
+
+      logger.info(
+        `🧪 Testing Bedrock account connection: ${account.name} (${accountId}), model: ${model}, credentialType: ${account.credentialType}`
+      )
+
+      // 设置 SSE 响应头
+      res.setHeader('Content-Type', 'text/event-stream')
+      res.setHeader('Cache-Control', 'no-cache')
+      res.setHeader('Connection', 'keep-alive')
+      res.setHeader('X-Accel-Buffering', 'no')
+      res.status(200)
+
+      // 发送 test_start 事件
+      res.write(`data: ${JSON.stringify({ type: 'test_start' })}\n\n`)
+
+      // 构造测试请求体（Bedrock 格式）
+      const bedrockPayload = {
+        anthropic_version: 'bedrock-2023-05-31',
+        max_tokens: 256,
+        messages: [
+          {
+            role: 'user',
+            content:
+              'Hello! Please respond with a simple greeting to confirm the connection is working. And tell me who are you?'
+          }
+        ]
+      }
+
+      // 获取 Bedrock 客户端
+      const region = account.region || bedrockRelayService.defaultRegion
+      const client = bedrockRelayService._getBedrockClient(region, account)
+
+      // 创建流式调用命令
+      const command = new InvokeModelWithResponseStreamCommand({
+        modelId: model,
+        body: JSON.stringify(bedrockPayload),
+        contentType: 'application/json',
+        accept: 'application/json'
+      })
+
+      logger.debug(`🌊 Bedrock test stream - model: ${model}, region: ${region}`)
+
+      const startTime = Date.now()
+      const response = await client.send(command)
+
+      // 处理流式响应
+      // let responseText = ''
+      for await (const chunk of response.body) {
+        if (chunk.chunk) {
+          const chunkData = JSON.parse(new TextDecoder().decode(chunk.chunk.bytes))
+
+          // 提取文本内容
+          if (chunkData.type === 'content_block_delta' && chunkData.delta?.text) {
+            const { text } = chunkData.delta
+            // responseText += text
+
+            // 发送 content 事件
+            res.write(`data: ${JSON.stringify({ type: 'content', text })}\n\n`)
+          }
+
+          // 检测错误
+          if (chunkData.type === 'error') {
+            throw new Error(chunkData.error?.message || 'Bedrock API error')
+          }
+        }
+      }
+
+      const duration = Date.now() - startTime
+      logger.info(`✅ Bedrock test completed - model: ${model}, duration: ${duration}ms`)
+
+      // 发送 message_stop 事件（前端兼容）
+      res.write(`data: ${JSON.stringify({ type: 'message_stop' })}\n\n`)
+
+      // 发送 test_complete 事件
+      res.write(`data: ${JSON.stringify({ type: 'test_complete', success: true })}\n\n`)
+
+      // 结束响应
+      res.end()
+
+      logger.info(`✅ Test request completed for Bedrock account: ${account.name}`)
+    } catch (error) {
+      logger.error(`❌ Test Bedrock account connection failed:`, error)
+
+      // 发送错误事件给前端
+      try {
+        // 检查响应流是否仍然可写
+        if (!res.writableEnded && !res.destroyed) {
+          if (!res.headersSent) {
+            res.setHeader('Content-Type', 'text/event-stream')
+            res.setHeader('Cache-Control', 'no-cache')
+            res.setHeader('Connection', 'keep-alive')
+            res.status(200)
+          }
+          const errorMsg = error.message || '测试失败'
+          res.write(`data: ${JSON.stringify({ type: 'error', error: errorMsg })}\n\n`)
+          res.end()
+        }
+      } catch (writeError) {
+        logger.error('Failed to write error to response stream:', writeError)
+      }
+
+      // 不再重新抛出错误，避免路由层再次处理
+      // throw error
+    }
+  }
+
   /**
    * 检查账户订阅是否过期
    * @param {Object} account - 账户对象

src/services/bedrockRelayService.js

@@ -6,6 +6,7 @@ const {
 const { fromEnv } = require('@aws-sdk/credential-providers')
 const logger = require('../utils/logger')
 const config = require('../../config/config')
+const userMessageQueueService = require('./userMessageQueueService')
 
 class BedrockRelayService {
   constructor() {
@@ -47,13 +48,17 @@ class BedrockRelayService {
         secretAccessKey: bedrockAccount.awsCredentials.secretAccessKey,
         sessionToken: bedrockAccount.awsCredentials.sessionToken
       }
+    } else if (bedrockAccount?.bearerToken) {
+      // Bearer Token 模式：AWS SDK >= 3.400.0 会自动检测环境变量
+      clientConfig.token = { token: bedrockAccount.bearerToken }
+      logger.debug(`🔑 使用 Bearer Token 认证 - 账户: ${bedrockAccount.name || 'unknown'}`)
     } else {
       // 检查是否有环境变量凭证
       if (process.env.AWS_ACCESS_KEY_ID && process.env.AWS_SECRET_ACCESS_KEY) {
         clientConfig.credentials = fromEnv()
       } else {
         throw new Error(
-          'AWS凭证未配置。请在Bedrock账户中配置AWS访问密钥，或设置环境变量AWS_ACCESS_KEY_ID和AWS_SECRET_ACCESS_KEY'
+          'AWS凭证未配置。请在Bedrock账户中配置AWS访问密钥或Bearer Token，或设置环境变量AWS_ACCESS_KEY_ID和AWS_SECRET_ACCESS_KEY'
         )
       }
     }
@@ -69,7 +74,68 @@ class BedrockRelayService {
 
   // 处理非流式请求
   async handleNonStreamRequest(requestBody, bedrockAccount = null) {
+    const accountId = bedrockAccount?.id
+    let queueLockAcquired = false
+    let queueRequestId = null
+
     try {
+      // 📬 用户消息队列处理
+      if (userMessageQueueService.isUserMessageRequest(requestBody)) {
+        // 校验 accountId 非空，避免空值污染队列锁键
+        if (!accountId || accountId === '') {
+          logger.error('❌ accountId missing for queue lock in Bedrock handleNonStreamRequest')
+          throw new Error('accountId missing for queue lock')
+        }
+        const queueResult = await userMessageQueueService.acquireQueueLock(accountId)
+        if (!queueResult.acquired && !queueResult.skipped) {
+          // 区分 Redis 后端错误和队列超时
+          const isBackendError = queueResult.error === 'queue_backend_error'
+          const errorCode = isBackendError ? 'QUEUE_BACKEND_ERROR' : 'QUEUE_TIMEOUT'
+          const errorType = isBackendError ? 'queue_backend_error' : 'queue_timeout'
+          const errorMessage = isBackendError
+            ? 'Queue service temporarily unavailable, please retry later'
+            : 'User message queue wait timeout, please retry later'
+          const statusCode = isBackendError ? 500 : 503
+
+          // 结构化性能日志，用于后续统计
+          logger.performance('user_message_queue_error', {
+            errorType,
+            errorCode,
+            accountId,
+            statusCode,
+            backendError: isBackendError ? queueResult.errorMessage : undefined
+          })
+
+          logger.warn(
+            `📬 User message queue ${errorType} for Bedrock account ${accountId}`,
+            isBackendError ? { backendError: queueResult.errorMessage } : {}
+          )
+          return {
+            statusCode,
+            headers: {
+              'Content-Type': 'application/json',
+              'x-user-message-queue-error': errorType
+            },
+            body: JSON.stringify({
+              type: 'error',
+              error: {
+                type: errorType,
+                code: errorCode,
+                message: errorMessage
+              }
+            }),
+            success: false
+          }
+        }
+        if (queueResult.acquired && !queueResult.skipped) {
+          queueLockAcquired = true
+          queueRequestId = queueResult.requestId
+          logger.debug(
+            `📬 User message queue lock acquired for Bedrock account ${accountId}, requestId: ${queueRequestId}`
+          )
+        }
+      }
+
       const modelId = this._selectModel(requestBody, bedrockAccount)
       const region = this._selectRegion(modelId, bedrockAccount)
       const client = this._getBedrockClient(region, bedrockAccount)
@@ -90,6 +156,23 @@ class BedrockRelayService {
       const response = await client.send(command)
       const duration = Date.now() - startTime
 
+      // 📬 请求已发送成功，立即释放队列锁（无需等待响应处理完成）
+      // 因为限流基于请求发送时刻计算（RPM），不是请求完成时刻
+      if (queueLockAcquired && queueRequestId && accountId) {
+        try {
+          await userMessageQueueService.releaseQueueLock(accountId, queueRequestId)
+          queueLockAcquired = false // 标记已释放，防止 finally 重复释放
+          logger.debug(
+            `📬 User message queue lock released early for Bedrock account ${accountId}, requestId: ${queueRequestId}`
+          )
+        } catch (releaseError) {
+          logger.error(
+            `❌ Failed to release user message queue lock early for Bedrock account ${accountId}:`,
+            releaseError.message
+          )
+        }
+      }
+
       // 解析响应
       const responseBody = JSON.parse(new TextDecoder().decode(response.body))
       const claudeResponse = this._convertFromBedrockFormat(responseBody)
@@ -106,12 +189,94 @@ class BedrockRelayService {
     } catch (error) {
       logger.error('❌ Bedrock非流式请求失败:', error)
       throw this._handleBedrockError(error)
+    } finally {
+      // 📬 释放用户消息队列锁（兜底，正常情况下已在请求发送后提前释放）
+      if (queueLockAcquired && queueRequestId && accountId) {
+        try {
+          await userMessageQueueService.releaseQueueLock(accountId, queueRequestId)
+          logger.debug(
+            `📬 User message queue lock released in finally for Bedrock account ${accountId}, requestId: ${queueRequestId}`
+          )
+        } catch (releaseError) {
+          logger.error(
+            `❌ Failed to release user message queue lock for Bedrock account ${accountId}:`,
+            releaseError.message
+          )
+        }
+      }
     }
   }
 
   // 处理流式请求
   async handleStreamRequest(requestBody, bedrockAccount = null, res) {
+    const accountId = bedrockAccount?.id
+    let queueLockAcquired = false
+    let queueRequestId = null
+
     try {
+      // 📬 用户消息队列处理
+      if (userMessageQueueService.isUserMessageRequest(requestBody)) {
+        // 校验 accountId 非空，避免空值污染队列锁键
+        if (!accountId || accountId === '') {
+          logger.error('❌ accountId missing for queue lock in Bedrock handleStreamRequest')
+          throw new Error('accountId missing for queue lock')
+        }
+        const queueResult = await userMessageQueueService.acquireQueueLock(accountId)
+        if (!queueResult.acquired && !queueResult.skipped) {
+          // 区分 Redis 后端错误和队列超时
+          const isBackendError = queueResult.error === 'queue_backend_error'
+          const errorCode = isBackendError ? 'QUEUE_BACKEND_ERROR' : 'QUEUE_TIMEOUT'
+          const errorType = isBackendError ? 'queue_backend_error' : 'queue_timeout'
+          const errorMessage = isBackendError
+            ? 'Queue service temporarily unavailable, please retry later'
+            : 'User message queue wait timeout, please retry later'
+          const statusCode = isBackendError ? 500 : 503
+
+          // 结构化性能日志，用于后续统计
+          logger.performance('user_message_queue_error', {
+            errorType,
+            errorCode,
+            accountId,
+            statusCode,
+            stream: true,
+            backendError: isBackendError ? queueResult.errorMessage : undefined
+          })
+
+          logger.warn(
+            `📬 User message queue ${errorType} for Bedrock account ${accountId} (stream)`,
+            isBackendError ? { backendError: queueResult.errorMessage } : {}
+          )
+          if (!res.headersSent) {
+            const existingConnection = res.getHeader ? res.getHeader('Connection') : null
+            res.writeHead(statusCode, {
+              'Content-Type': 'text/event-stream',
+              'Cache-Control': 'no-cache',
+              Connection: existingConnection || 'keep-alive',
+              'x-user-message-queue-error': errorType
+            })
+          }
+          const errorEvent = `event: error\ndata: ${JSON.stringify({
+            type: 'error',
+            error: {
+              type: errorType,
+              code: errorCode,
+              message: errorMessage
+            }
+          })}\n\n`
+          res.write(errorEvent)
+          res.write('data: [DONE]\n\n')
+          res.end()
+          return { success: false, error: errorType }
+        }
+        if (queueResult.acquired && !queueResult.skipped) {
+          queueLockAcquired = true
+          queueRequestId = queueResult.requestId
+          logger.debug(
+            `📬 User message queue lock acquired for Bedrock account ${accountId} (stream), requestId: ${queueRequestId}`
+          )
+        }
+      }
+
       const modelId = this._selectModel(requestBody, bedrockAccount)
       const region = this._selectRegion(modelId, bedrockAccount)
       const client = this._getBedrockClient(region, bedrockAccount)
@@ -131,11 +296,35 @@ class BedrockRelayService {
       const startTime = Date.now()
       const response = await client.send(command)
 
+      // 📬 请求已发送成功，立即释放队列锁（无需等待响应处理完成）
+      // 因为限流基于请求发送时刻计算（RPM），不是请求完成时刻
+      if (queueLockAcquired && queueRequestId && accountId) {
+        try {
+          await userMessageQueueService.releaseQueueLock(accountId, queueRequestId)
+          queueLockAcquired = false // 标记已释放，防止 finally 重复释放
+          logger.debug(
+            `📬 User message queue lock released early for Bedrock stream account ${accountId}, requestId: ${queueRequestId}`
+          )
+        } catch (releaseError) {
+          logger.error(
+            `❌ Failed to release user message queue lock early for Bedrock stream account ${accountId}:`,
+            releaseError.message
+          )
+        }
+      }
+
       // 设置SSE响应头
+      // ⚠️ 关键修复：尊重 auth.js 提前设置的 Connection: close
+      const existingConnection = res.getHeader ? res.getHeader('Connection') : null
+      if (existingConnection) {
+        logger.debug(
+          `🔌 [Bedrock Stream] Preserving existing Connection header: ${existingConnection}`
+        )
+      }
       res.writeHead(200, {
         'Content-Type': 'text/event-stream',
         'Cache-Control': 'no-cache',
-        Connection: 'keep-alive',
+        Connection: existingConnection || 'keep-alive',
         'Access-Control-Allow-Origin': '*',
         'Access-Control-Allow-Headers': 'Content-Type, Authorization'
       })
@@ -154,8 +343,8 @@ class BedrockRelayService {
             res.write(`event: ${claudeEvent.type}\n`)
             res.write(`data: ${JSON.stringify(claudeEvent.data)}\n\n`)
 
-            // 提取使用统计
-            if (claudeEvent.type === 'message_stop' && claudeEvent.data.usage) {
+            // 提取使用统计 (usage is reported in message_delta per Claude API spec)
+            if (claudeEvent.type === 'message_delta' && claudeEvent.data.usage) {
               totalUsage = claudeEvent.data.usage
             }
 
@@ -191,6 +380,21 @@ class BedrockRelayService {
       res.end()
 
       throw this._handleBedrockError(error)
+    } finally {
+      // 📬 释放用户消息队列锁（兜底，正常情况下已在请求发送后提前释放）
+      if (queueLockAcquired && queueRequestId && accountId) {
+        try {
+          await userMessageQueueService.releaseQueueLock(accountId, queueRequestId)
+          logger.debug(
+            `📬 User message queue lock released in finally for Bedrock stream account ${accountId}, requestId: ${queueRequestId}`
+          )
+        } catch (releaseError) {
+          logger.error(
+            `❌ Failed to release user message queue lock for Bedrock stream account ${accountId}:`,
+            releaseError.message
+          )
+        }
+      }
     }
   }
 
@@ -231,6 +435,18 @@ class BedrockRelayService {
   _mapToBedrockModel(modelName) {
     // 标准Claude模型名到Bedrock模型名的映射表
     const modelMapping = {
+      // Claude 4.5 Opus
+      'claude-opus-4-5': 'us.anthropic.claude-opus-4-5-20251101-v1:0',
+      'claude-opus-4-5-20251101': 'us.anthropic.claude-opus-4-5-20251101-v1:0',
+
+      // Claude 4.5 Sonnet
+      'claude-sonnet-4-5': 'us.anthropic.claude-sonnet-4-5-20250929-v1:0',
+      'claude-sonnet-4-5-20250929': 'us.anthropic.claude-sonnet-4-5-20250929-v1:0',
+
+      // Claude 4.5 Haiku
+      'claude-haiku-4-5': 'us.anthropic.claude-haiku-4-5-20251001-v1:0',
+      'claude-haiku-4-5-20251001': 'us.anthropic.claude-haiku-4-5-20251001-v1:0',
+
       // Claude Sonnet 4
       'claude-sonnet-4': 'us.anthropic.claude-sonnet-4-20250514-v1:0',
       'claude-sonnet-4-20250514': 'us.anthropic.claude-sonnet-4-20250514-v1:0',
@@ -360,14 +576,28 @@ class BedrockRelayService {
       return {
         type: 'message_start',
         data: {
-          type: 'message',
-          id: `msg_${Date.now()}_bedrock`,
-          role: 'assistant',
-          content: [],
-          model: this.defaultModel,
-          stop_reason: null,
-          stop_sequence: null,
-          usage: bedrockChunk.message?.usage || { input_tokens: 0, output_tokens: 0 }
+          type: 'message_start',
+          message: {
+            id: `msg_${Date.now()}_bedrock`,
+            type: 'message',
+            role: 'assistant',
+            content: [],
+            model: this.defaultModel,
+            stop_reason: null,
+            stop_sequence: null,
+            usage: bedrockChunk.message?.usage || { input_tokens: 0, output_tokens: 0 }
+          }
+        }
+      }
+    }
+
+    if (bedrockChunk.type === 'content_block_start') {
+      return {
+        type: 'content_block_start',
+        data: {
+          type: 'content_block_start',
+          index: bedrockChunk.index || 0,
+          content_block: bedrockChunk.content_block || { type: 'text', text: '' }
         }
       }
     }
@@ -376,16 +606,28 @@ class BedrockRelayService {
       return {
         type: 'content_block_delta',
         data: {
+          type: 'content_block_delta',
           index: bedrockChunk.index || 0,
           delta: bedrockChunk.delta || {}
         }
       }
     }
 
+    if (bedrockChunk.type === 'content_block_stop') {
+      return {
+        type: 'content_block_stop',
+        data: {
+          type: 'content_block_stop',
+          index: bedrockChunk.index || 0
+        }
+      }
+    }
+
     if (bedrockChunk.type === 'message_delta') {
       return {
         type: 'message_delta',
         data: {
+          type: 'message_delta',
           delta: bedrockChunk.delta || {},
           usage: bedrockChunk.usage || {}
         }
@@ -396,7 +638,7 @@ class BedrockRelayService {
       return {
         type: 'message_stop',
         data: {
-          usage: bedrockChunk.usage || {}
+          type: 'message_stop'
         }
       }
     }

src/services/ccrRelayService.js

@@ -3,6 +3,8 @@ const ccrAccountService = require('./ccrAccountService')
 const logger = require('../utils/logger')
 const config = require('../../config/config')
 const { parseVendorPrefixedModel } = require('../utils/modelHelper')
+const userMessageQueueService = require('./userMessageQueueService')
+const { isStreamWritable } = require('../utils/streamHelper')
 
 class CcrRelayService {
   constructor() {
@@ -21,8 +23,67 @@ class CcrRelayService {
   ) {
     let abortController = null
     let account = null
+    let queueLockAcquired = false
+    let queueRequestId = null
 
     try {
+      // 📬 用户消息队列处理
+      if (userMessageQueueService.isUserMessageRequest(requestBody)) {
+        // 校验 accountId 非空，避免空值污染队列锁键
+        if (!accountId || accountId === '') {
+          logger.error('❌ accountId missing for queue lock in CCR relayRequest')
+          throw new Error('accountId missing for queue lock')
+        }
+        const queueResult = await userMessageQueueService.acquireQueueLock(accountId)
+        if (!queueResult.acquired && !queueResult.skipped) {
+          // 区分 Redis 后端错误和队列超时
+          const isBackendError = queueResult.error === 'queue_backend_error'
+          const errorCode = isBackendError ? 'QUEUE_BACKEND_ERROR' : 'QUEUE_TIMEOUT'
+          const errorType = isBackendError ? 'queue_backend_error' : 'queue_timeout'
+          const errorMessage = isBackendError
+            ? 'Queue service temporarily unavailable, please retry later'
+            : 'User message queue wait timeout, please retry later'
+          const statusCode = isBackendError ? 500 : 503
+
+          // 结构化性能日志，用于后续统计
+          logger.performance('user_message_queue_error', {
+            errorType,
+            errorCode,
+            accountId,
+            statusCode,
+            backendError: isBackendError ? queueResult.errorMessage : undefined
+          })
+
+          logger.warn(
+            `📬 User message queue ${errorType} for CCR account ${accountId}`,
+            isBackendError ? { backendError: queueResult.errorMessage } : {}
+          )
+          return {
+            statusCode,
+            headers: {
+              'Content-Type': 'application/json',
+              'x-user-message-queue-error': errorType
+            },
+            body: JSON.stringify({
+              type: 'error',
+              error: {
+                type: errorType,
+                code: errorCode,
+                message: errorMessage
+              }
+            }),
+            accountId
+          }
+        }
+        if (queueResult.acquired && !queueResult.skipped) {
+          queueLockAcquired = true
+          queueRequestId = queueResult.requestId
+          logger.debug(
+            `📬 User message queue lock acquired for CCR account ${accountId}, requestId: ${queueRequestId}`
+          )
+        }
+      }
+
       // 获取账户信息
       account = await ccrAccountService.getAccount(accountId)
       if (!account) {
@@ -162,6 +223,23 @@ class CcrRelayService {
       )
       const response = await axios(requestConfig)
 
+      // 📬 请求已发送成功，立即释放队列锁（无需等待响应处理完成）
+      // 因为 Claude API 限流基于请求发送时刻计算（RPM），不是请求完成时刻
+      if (queueLockAcquired && queueRequestId && accountId) {
+        try {
+          await userMessageQueueService.releaseQueueLock(accountId, queueRequestId)
+          queueLockAcquired = false // 标记已释放，防止 finally 重复释放
+          logger.debug(
+            `📬 User message queue lock released early for CCR account ${accountId}, requestId: ${queueRequestId}`
+          )
+        } catch (releaseError) {
+          logger.error(
+            `❌ Failed to release user message queue lock early for CCR account ${accountId}:`,
+            releaseError.message
+          )
+        }
+      }
+
       // 移除监听器（请求成功完成）
       if (clientRequest) {
         clientRequest.removeListener('close', handleClientDisconnect)
@@ -233,6 +311,21 @@ class CcrRelayService {
       )
 
       throw error
+    } finally {
+      // 📬 释放用户消息队列锁（兜底，正常情况下已在请求发送后提前释放）
+      if (queueLockAcquired && queueRequestId && accountId) {
+        try {
+          await userMessageQueueService.releaseQueueLock(accountId, queueRequestId)
+          logger.debug(
+            `📬 User message queue lock released in finally for CCR account ${accountId}, requestId: ${queueRequestId}`
+          )
+        } catch (releaseError) {
+          logger.error(
+            `❌ Failed to release user message queue lock for CCR account ${accountId}:`,
+            releaseError.message
+          )
+        }
+      }
     }
   }
 
@@ -248,7 +341,77 @@ class CcrRelayService {
     options = {}
   ) {
     let account = null
+    let queueLockAcquired = false
+    let queueRequestId = null
+
     try {
+      // 📬 用户消息队列处理
+      if (userMessageQueueService.isUserMessageRequest(requestBody)) {
+        // 校验 accountId 非空，避免空值污染队列锁键
+        if (!accountId || accountId === '') {
+          logger.error(
+            '❌ accountId missing for queue lock in CCR relayStreamRequestWithUsageCapture'
+          )
+          throw new Error('accountId missing for queue lock')
+        }
+        const queueResult = await userMessageQueueService.acquireQueueLock(accountId)
+        if (!queueResult.acquired && !queueResult.skipped) {
+          // 区分 Redis 后端错误和队列超时
+          const isBackendError = queueResult.error === 'queue_backend_error'
+          const errorCode = isBackendError ? 'QUEUE_BACKEND_ERROR' : 'QUEUE_TIMEOUT'
+          const errorType = isBackendError ? 'queue_backend_error' : 'queue_timeout'
+          const errorMessage = isBackendError
+            ? 'Queue service temporarily unavailable, please retry later'
+            : 'User message queue wait timeout, please retry later'
+          const statusCode = isBackendError ? 500 : 503
+
+          // 结构化性能日志，用于后续<E5908E><E7BBAD>计
+          logger.performance('user_message_queue_error', {
+            errorType,
+            errorCode,
+            accountId,
+            statusCode,
+            stream: true,
+            backendError: isBackendError ? queueResult.errorMessage : undefined
+          })
+
+          logger.warn(
+            `📬 User message queue ${errorType} for CCR account ${accountId} (stream)`,
+            isBackendError ? { backendError: queueResult.errorMessage } : {}
+          )
+          if (!responseStream.headersSent) {
+            const existingConnection = responseStream.getHeader
+              ? responseStream.getHeader('Connection')
+              : null
+            responseStream.writeHead(statusCode, {
+              'Content-Type': 'text/event-stream',
+              'Cache-Control': 'no-cache',
+              Connection: existingConnection || 'keep-alive',
+              'x-user-message-queue-error': errorType
+            })
+          }
+          const errorEvent = `event: error\ndata: ${JSON.stringify({
+            type: 'error',
+            error: {
+              type: errorType,
+              code: errorCode,
+              message: errorMessage
+            }
+          })}\n\n`
+          responseStream.write(errorEvent)
+          responseStream.write('data: [DONE]\n\n')
+          responseStream.end()
+          return
+        }
+        if (queueResult.acquired && !queueResult.skipped) {
+          queueLockAcquired = true
+          queueRequestId = queueResult.requestId
+          logger.debug(
+            `📬 User message queue lock acquired for CCR account ${accountId} (stream), requestId: ${queueRequestId}`
+          )
+        }
+      }
+
       // 获取账户信息
       account = await ccrAccountService.getAccount(accountId)
       if (!account) {
@@ -296,14 +459,53 @@ class CcrRelayService {
         accountId,
         usageCallback,
         streamTransformer,
-        options
+        options,
+        // 📬 回调：在收到响应头时释放队列锁
+        async () => {
+          if (queueLockAcquired && queueRequestId && accountId) {
+            try {
+              await userMessageQueueService.releaseQueueLock(accountId, queueRequestId)
+              queueLockAcquired = false // 标记已释放，防止 finally 重复释放
+              logger.debug(
+                `📬 User message queue lock released early for CCR stream account ${accountId}, requestId: ${queueRequestId}`
+              )
+            } catch (releaseError) {
+              logger.error(
+                `❌ Failed to release user message queue lock early for CCR stream account ${accountId}:`,
+                releaseError.message
+              )
+            }
+          }
+        }
       )
 
       // 更新最后使用时间
       await this._updateLastUsedTime(accountId)
     } catch (error) {
-      logger.error(`❌ CCR stream relay failed (Account: ${account?.name || accountId}):`, error)
+      // 客户端主动断开连接是正常情况，使用 INFO 级别
+      if (error.message === 'Client disconnected') {
+        logger.info(
+          `🔌 CCR stream relay ended: Client disconnected (Account: ${account?.name || accountId})`
+        )
+      } else {
+        logger.error(`❌ CCR stream relay failed (Account: ${account?.name || accountId}):`, error)
+      }
       throw error
+    } finally {
+      // 📬 释放用户消息队列锁（兜底，正常情况下已在收到响应头后提前释放）
+      if (queueLockAcquired && queueRequestId && accountId) {
+        try {
+          await userMessageQueueService.releaseQueueLock(accountId, queueRequestId)
+          logger.debug(
+            `📬 User message queue lock released in finally for CCR stream account ${accountId}, requestId: ${queueRequestId}`
+          )
+        } catch (releaseError) {
+          logger.error(
+            `❌ Failed to release user message queue lock for CCR stream account ${accountId}:`,
+            releaseError.message
+          )
+        }
+      }
     }
   }
 
@@ -317,7 +519,8 @@ class CcrRelayService {
     accountId,
     usageCallback,
     streamTransformer = null,
-    requestOptions = {}
+    requestOptions = {},
+    onResponseHeaderReceived = null
   ) {
     return new Promise((resolve, reject) => {
       let aborted = false
@@ -380,8 +583,11 @@ class CcrRelayService {
       // 发送请求
       const request = axios(requestConfig)
 
+      // 注意：使用 .then(async ...) 模式处理响应
+      // - 内部的 releaseQueueLock 有独立的 try-catch，不会导致未捕获异常
+      // - queueLockAcquired = false 的赋值会在 finally 执行前完成（JS 单线程保证）
       request
-        .then((response) => {
+        .then(async (response) => {
           logger.debug(`🌊 CCR stream response status: ${response.status}`)
 
           // 错误响应处理
@@ -404,10 +610,13 @@ class CcrRelayService {
 
             // 设置错误响应的状态码和响应头
             if (!responseStream.headersSent) {
+              const existingConnection = responseStream.getHeader
+                ? responseStream.getHeader('Connection')
+                : null
               const errorHeaders = {
                 'Content-Type': response.headers['content-type'] || 'application/json',
                 'Cache-Control': 'no-cache',
-                Connection: 'keep-alive'
+                Connection: existingConnection || 'keep-alive'
               }
               // 避免 Transfer-Encoding 冲突，让 Express 自动处理
               delete errorHeaders['Transfer-Encoding']
@@ -417,13 +626,13 @@ class CcrRelayService {
 
             // 直接透传错误数据，不进行包装
             response.data.on('data', (chunk) => {
-              if (!responseStream.destroyed) {
+              if (isStreamWritable(responseStream)) {
                 responseStream.write(chunk)
               }
             })
 
             response.data.on('end', () => {
-              if (!responseStream.destroyed) {
+              if (isStreamWritable(responseStream)) {
                 responseStream.end()
               }
               resolve() // 不抛出异常，正常完成流处理
@@ -431,6 +640,19 @@ class CcrRelayService {
             return
           }
 
+          // 📬 收到成功响应头（HTTP 200），调用回调释放队列锁
+          // 此时请求已被 Claude API 接受并计入 RPM 配额，无需等待响应完成
+          if (onResponseHeaderReceived && typeof onResponseHeaderReceived === 'function') {
+            try {
+              await onResponseHeaderReceived()
+            } catch (callbackError) {
+              logger.error(
+                `❌ Failed to execute onResponseHeaderReceived callback for CCR stream account ${accountId}:`,
+                callbackError.message
+              )
+            }
+          }
+
           // 成功响应，检查并移除错误状态
           ccrAccountService.isAccountRateLimited(accountId).then((isRateLimited) => {
             if (isRateLimited) {
@@ -444,11 +666,20 @@ class CcrRelayService {
           })
 
           // 设置响应头
+          // ⚠️ 关键修复：尊重 auth.js 提前设置的 Connection: close
           if (!responseStream.headersSent) {
+            const existingConnection = responseStream.getHeader
+              ? responseStream.getHeader('Connection')
+              : null
+            if (existingConnection) {
+              logger.debug(
+                `🔌 [CCR Stream] Preserving existing Connection header: ${existingConnection}`
+              )
+            }
             const headers = {
               'Content-Type': 'text/event-stream',
               'Cache-Control': 'no-cache',
-              Connection: 'keep-alive',
+              Connection: existingConnection || 'keep-alive',
               'Access-Control-Allow-Origin': '*',
               'Access-Control-Allow-Headers': 'Cache-Control'
             }
@@ -487,12 +718,17 @@ class CcrRelayService {
                   }
 
                   // 写入到响应流
-                  if (outputLine && !responseStream.destroyed) {
+                  if (outputLine && isStreamWritable(responseStream)) {
                     responseStream.write(`${outputLine}\n`)
+                  } else if (outputLine) {
+                    // 客户端连接已断开，记录警告
+                    logger.warn(
+                      `⚠️ [CCR] Client disconnected during stream, skipping data for account: ${accountId}`
+                    )
                   }
                 } else {
                   // 空行也需要传递
-                  if (!responseStream.destroyed) {
+                  if (isStreamWritable(responseStream)) {
                     responseStream.write('\n')
                   }
                 }
@@ -503,10 +739,6 @@ class CcrRelayService {
           })
 
           response.data.on('end', () => {
-            if (!responseStream.destroyed) {
-              responseStream.end()
-            }
-
             // 如果收集到使用统计数据，调用回调
             if (usageCallback && Object.keys(collectedUsage).length > 0) {
               try {
@@ -518,12 +750,26 @@ class CcrRelayService {
               }
             }
 
-            resolve()
+            if (isStreamWritable(responseStream)) {
+              // 等待数据完全 flush 到客户端后再 resolve
+              responseStream.end(() => {
+                logger.debug(
+                  `🌊 CCR stream response completed and flushed | bytesWritten: ${responseStream.bytesWritten || 'unknown'}`
+                )
+                resolve()
+              })
+            } else {
+              // 连接已断开，记录警告
+              logger.warn(
+                `⚠️ [CCR] Client disconnected before stream end, data may not have been received | account: ${accountId}`
+              )
+              resolve()
+            }
           })
 
           response.data.on('error', (err) => {
             logger.error('❌ Stream data error:', err)
-            if (!responseStream.destroyed) {
+            if (isStreamWritable(responseStream)) {
               responseStream.end()
             }
             reject(err)
@@ -555,7 +801,7 @@ class CcrRelayService {
             }
           }
 
-          if (!responseStream.destroyed) {
+          if (isStreamWritable(responseStream)) {
             responseStream.write(`data: ${JSON.stringify(errorResponse)}\n\n`)
             responseStream.end()
           }

src/services/claudeAccountService.js

@@ -16,6 +16,22 @@ const {
 const tokenRefreshService = require('./tokenRefreshService')
 const LRUCache = require('../utils/lruCache')
 const { formatDateWithTimezone, getISOStringWithTimezone } = require('../utils/dateHelper')
+const { isOpus45OrNewer } = require('../utils/modelHelper')
+
+/**
+ * Check if account is Pro (not Max)
+ * Compatible with both API real-time data (hasClaudePro) and local config (accountType)
+ * @param {Object} info - Subscription info object
+ * @returns {boolean}
+ */
+function isProAccount(info) {
+  // API real-time status takes priority
+  if (info.hasClaudePro === true && info.hasClaudeMax !== true) {
+    return true
+  }
+  // Local configured account type
+  return info.accountType === 'claude_pro'
+}
 
 class ClaudeAccountService {
   constructor() {
@@ -75,7 +91,9 @@ class ClaudeAccountService {
       useUnifiedClientId = false, // 是否使用统一的客户端标识
       unifiedClientId = '', // 统一的客户端标识
       expiresAt = null, // 账户订阅到期时间
-      extInfo = null // 额外扩展信息
+      extInfo = null, // 额外扩展信息
+      maxConcurrency = 0, // 账户级用户消息串行队列：0=使用全局配置，>0=强制启用串行
+      interceptWarmup = false // 拦截预热请求（标题生成、Warmup等）
     } = options
 
     const accountId = uuidv4()
@@ -120,7 +138,11 @@ class ClaudeAccountService {
         // 账户订阅到期时间
         subscriptionExpiresAt: expiresAt || '',
         // 扩展信息
-        extInfo: normalizedExtInfo ? JSON.stringify(normalizedExtInfo) : ''
+        extInfo: normalizedExtInfo ? JSON.stringify(normalizedExtInfo) : '',
+        // 账户级用户消息串行队列限制
+        maxConcurrency: maxConcurrency.toString(),
+        // 拦截预热请求
+        interceptWarmup: interceptWarmup.toString()
       }
     } else {
       // 兼容旧格式
@@ -152,7 +174,11 @@ class ClaudeAccountService {
         // 账户订阅到期时间
         subscriptionExpiresAt: expiresAt || '',
         // 扩展信息
-        extInfo: normalizedExtInfo ? JSON.stringify(normalizedExtInfo) : ''
+        extInfo: normalizedExtInfo ? JSON.stringify(normalizedExtInfo) : '',
+        // 账户级用户消息串行队列限制
+        maxConcurrency: maxConcurrency.toString(),
+        // 拦截预热请求
+        interceptWarmup: interceptWarmup.toString()
       }
     }
 
@@ -200,7 +226,8 @@ class ClaudeAccountService {
       useUnifiedUserAgent,
       useUnifiedClientId,
       unifiedClientId,
-      extInfo: normalizedExtInfo
+      extInfo: normalizedExtInfo,
+      interceptWarmup
     }
   }
 
@@ -558,7 +585,11 @@ class ClaudeAccountService {
             // 添加停止原因
             stoppedReason: account.stoppedReason || null,
             // 扩展信息
-            extInfo: parsedExtInfo
+            extInfo: parsedExtInfo,
+            // 账户级用户消息串行队列限制
+            maxConcurrency: parseInt(account.maxConcurrency || '0', 10),
+            // 拦截预热请求
+            interceptWarmup: account.interceptWarmup === 'true'
           }
         })
       )
@@ -650,7 +681,9 @@ class ClaudeAccountService {
         'useUnifiedClientId',
         'unifiedClientId',
         'subscriptionExpiresAt',
-        'extInfo'
+        'extInfo',
+        'maxConcurrency',
+        'interceptWarmup'
       ]
       const updatedData = { ...accountData }
       let shouldClearAutoStopFields = false
@@ -665,7 +698,7 @@ class ClaudeAccountService {
             updatedData[field] = this._encryptSensitiveData(value)
           } else if (field === 'proxy') {
             updatedData[field] = value ? JSON.stringify(value) : ''
-          } else if (field === 'priority') {
+          } else if (field === 'priority' || field === 'maxConcurrency') {
             updatedData[field] = value.toString()
           } else if (field === 'subscriptionInfo') {
             // 处理订阅信息更新
@@ -852,31 +885,39 @@ class ClaudeAccountService {
           !this.isSubscriptionExpired(account)
       )
 
-      // 如果请求的是 Opus 模型，过滤掉 Pro 和 Free 账号
+      // Filter Opus models based on account type and model version
       if (modelName && modelName.toLowerCase().includes('opus')) {
+        const isNewOpus = isOpus45OrNewer(modelName)
+
         activeAccounts = activeAccounts.filter((account) => {
-          // 检查账号的订阅信息
           if (account.subscriptionInfo) {
             try {
               const info = JSON.parse(account.subscriptionInfo)
-              // Pro 和 Free 账号不支持 Opus
-              if (info.hasClaudePro === true && info.hasClaudeMax !== true) {
-                return false // Claude Pro 不支持 Opus
+
+              // Free account: does not support any Opus model
+              if (info.accountType === 'free') {
+                return false
               }
-              if (info.accountType === 'claude_pro' || info.accountType === 'claude_free') {
-                return false // 明确标记为 Pro 或 Free 的账号不支持
+
+              // Pro account: only supports Opus 4.5+
+              if (isProAccount(info)) {
+                return isNewOpus
               }
+
+              // Max account: supports all Opus versions
+              return true
             } catch (e) {
-              // 解析失败，假设为旧数据，默认支持（兼容旧数据为 Max）
+              // Parse failed, assume legacy data (Max), default support
               return true
             }
           }
-          // 没有订阅信息的账号，默认当作支持（兼容旧数据）
+          // Account without subscription info, default to supported (legacy data compatibility)
           return true
         })
 
         if (activeAccounts.length === 0) {
-          throw new Error('No Claude accounts available that support Opus model')
+          const modelDesc = isNewOpus ? 'Opus 4.5+' : 'legacy Opus (requires Max subscription)'
+          throw new Error(`No Claude accounts available that support ${modelDesc} model`)
         }
       }
 
@@ -970,31 +1011,39 @@ class ClaudeAccountService {
           !this.isSubscriptionExpired(account)
       )
 
-      // 如果请求的是 Opus 模型，过滤掉 Pro 和 Free 账号
+      // Filter Opus models based on account type and model version
       if (modelName && modelName.toLowerCase().includes('opus')) {
+        const isNewOpus = isOpus45OrNewer(modelName)
+
         sharedAccounts = sharedAccounts.filter((account) => {
-          // 检查账号的订阅信息
           if (account.subscriptionInfo) {
             try {
               const info = JSON.parse(account.subscriptionInfo)
-              // Pro 和 Free 账号不支持 Opus
-              if (info.hasClaudePro === true && info.hasClaudeMax !== true) {
-                return false // Claude Pro 不支持 Opus
+
+              // Free account: does not support any Opus model
+              if (info.accountType === 'free') {
+                return false
               }
-              if (info.accountType === 'claude_pro' || info.accountType === 'claude_free') {
-                return false // 明确标记为 Pro 或 Free 的账号不支持
+
+              // Pro account: only supports Opus 4.5+
+              if (isProAccount(info)) {
+                return isNewOpus
               }
+
+              // Max account: supports all Opus versions
+              return true
             } catch (e) {
-              // 解析失败，假设为旧数据，默认支持（兼容旧数据为 Max）
+              // Parse failed, assume legacy data (Max), default support
               return true
             }
           }
-          // 没有订阅信息的账号，默认当作支持（兼容旧数据）
+          // Account without subscription info, default to supported (legacy data compatibility)
           return true
         })
 
         if (sharedAccounts.length === 0) {
-          throw new Error('No shared Claude accounts available that support Opus model')
+          const modelDesc = isNewOpus ? 'Opus 4.5+' : 'legacy Opus (requires Max subscription)'
+          throw new Error(`No shared Claude accounts available that support ${modelDesc} model`)
         }
       }
 

src/services/claudeConsoleAccountService.js

@@ -67,7 +67,9 @@ class ClaudeConsoleAccountService {
       schedulable = true, // 是否可被调度
       dailyQuota = 0, // 每日额度限制（美元），0表示不限制
       quotaResetTime = '00:00', // 额度重置时间（HH:mm格式）
-      maxConcurrentTasks = 0 // 最大并发任务数，0表示无限制
+      maxConcurrentTasks = 0, // 最大并发任务数，0表示无限制
+      disableAutoProtection = false, // 是否关闭自动防护（429/401/400/529 不自动禁用）
+      interceptWarmup = false // 拦截预热请求（标题生成、Warmup等）
     } = options
 
     // 验证必填字段
@@ -115,7 +117,9 @@ class ClaudeConsoleAccountService {
       lastResetDate: redis.getDateStringInTimezone(), // 最后重置日期（按配置时区）
       quotaResetTime, // 额度重置时间
       quotaStoppedAt: '', // 因额度停用的时间
-      maxConcurrentTasks: maxConcurrentTasks.toString() // 最大并发任务数，0表示无限制
+      maxConcurrentTasks: maxConcurrentTasks.toString(), // 最大并发任务数，0表示无限制
+      disableAutoProtection: disableAutoProtection.toString(), // 关闭自动防护
+      interceptWarmup: interceptWarmup.toString() // 拦截预热请求
     }
 
     const client = redis.getClientSafe()
@@ -153,6 +157,8 @@ class ClaudeConsoleAccountService {
       quotaResetTime,
       quotaStoppedAt: null,
       maxConcurrentTasks, // 新增：返回并发限制配置
+      disableAutoProtection, // 新增：返回自动防护开关
+      interceptWarmup, // 新增：返回预热请求拦截开关
       activeTaskCount: 0 // 新增：新建账户当前并发数为0
     }
   }
@@ -213,7 +219,10 @@ class ClaudeConsoleAccountService {
 
             // 并发控制相关
             maxConcurrentTasks: parseInt(accountData.maxConcurrentTasks) || 0,
-            activeTaskCount
+            activeTaskCount,
+            disableAutoProtection: accountData.disableAutoProtection === 'true',
+            // 拦截预热请求
+            interceptWarmup: accountData.interceptWarmup === 'true'
           })
         }
       }
@@ -259,6 +268,7 @@ class ClaudeConsoleAccountService {
     }
     accountData.isActive = accountData.isActive === 'true'
     accountData.schedulable = accountData.schedulable !== 'false' // 默认为true
+    accountData.disableAutoProtection = accountData.disableAutoProtection === 'true'
 
     if (accountData.proxy) {
       accountData.proxy = JSON.parse(accountData.proxy)
@@ -367,6 +377,12 @@ class ClaudeConsoleAccountService {
       if (updates.maxConcurrentTasks !== undefined) {
         updatedData.maxConcurrentTasks = updates.maxConcurrentTasks.toString()
       }
+      if (updates.disableAutoProtection !== undefined) {
+        updatedData.disableAutoProtection = updates.disableAutoProtection.toString()
+      }
+      if (updates.interceptWarmup !== undefined) {
+        updatedData.interceptWarmup = updates.interceptWarmup.toString()
+      }
 
       // ✅ 直接保存 subscriptionExpiresAt（如果提供）
       // Claude Console 没有 token 刷新逻辑，不会覆盖此字段
@@ -1510,6 +1526,71 @@ class ClaudeConsoleAccountService {
     const expiryDate = new Date(account.subscriptionExpiresAt)
     return expiryDate <= new Date()
   }
+
+  // 🚫 标记账户的 count_tokens 端点不可用
+  async markCountTokensUnavailable(accountId) {
+    try {
+      const client = redis.getClientSafe()
+      const accountKey = `${this.ACCOUNT_KEY_PREFIX}${accountId}`
+
+      // 检查账户是否存在
+      const exists = await client.exists(accountKey)
+      if (!exists) {
+        logger.warn(
+          `⚠️ Cannot mark count_tokens unavailable for non-existent account: ${accountId}`
+        )
+        return { success: false, reason: 'Account not found' }
+      }
+
+      await client.hset(accountKey, {
+        countTokensUnavailable: 'true',
+        countTokensUnavailableAt: new Date().toISOString()
+      })
+
+      logger.info(
+        `🚫 Marked count_tokens endpoint as unavailable for Claude Console account: ${accountId}`
+      )
+      return { success: true }
+    } catch (error) {
+      logger.error(`❌ Failed to mark count_tokens unavailable for account ${accountId}:`, error)
+      throw error
+    }
+  }
+
+  // ✅ 移除账户的 count_tokens 不可用标记
+  async removeCountTokensUnavailable(accountId) {
+    try {
+      const client = redis.getClientSafe()
+      const accountKey = `${this.ACCOUNT_KEY_PREFIX}${accountId}`
+
+      await client.hdel(accountKey, 'countTokensUnavailable', 'countTokensUnavailableAt')
+
+      logger.info(
+        `✅ Removed count_tokens unavailable mark for Claude Console account: ${accountId}`
+      )
+      return { success: true }
+    } catch (error) {
+      logger.error(
+        `❌ Failed to remove count_tokens unavailable mark for account ${accountId}:`,
+        error
+      )
+      throw error
+    }
+  }
+
+  // 🔍 检查账户的 count_tokens 端点是否不可用
+  async isCountTokensUnavailable(accountId) {
+    try {
+      const client = redis.getClientSafe()
+      const accountKey = `${this.ACCOUNT_KEY_PREFIX}${accountId}`
+
+      const value = await client.hget(accountKey, 'countTokensUnavailable')
+      return value === 'true'
+    } catch (error) {
+      logger.error(`❌ Failed to check count_tokens availability for account ${accountId}:`, error)
+      return false // 出错时默认返回可用，避免误阻断
+    }
+  }
 }
 
 module.exports = new ClaudeConsoleAccountService()

src/services/claudeConsoleRelayService.js

@@ -9,6 +9,9 @@ const {
   sanitizeErrorMessage,
   isAccountDisabledError
 } = require('../utils/errorSanitizer')
+const userMessageQueueService = require('./userMessageQueueService')
+const { isStreamWritable } = require('../utils/streamHelper')
+const { filterForClaude } = require('../utils/headerFilter')
 
 class ClaudeConsoleRelayService {
   constructor() {
@@ -29,14 +32,76 @@ class ClaudeConsoleRelayService {
     let account = null
     const requestId = uuidv4() // 用于并发追踪
     let concurrencyAcquired = false
+    let queueLockAcquired = false
+    let queueRequestId = null
 
     try {
+      // 📬 用户消息队列处理：如果是用户消息请求，需要获取队列锁
+      if (userMessageQueueService.isUserMessageRequest(requestBody)) {
+        // 校验 accountId 非空，避免空值污染队列锁键
+        if (!accountId || accountId === '') {
+          logger.error('❌ accountId missing for queue lock in console relayRequest')
+          throw new Error('accountId missing for queue lock')
+        }
+        const queueResult = await userMessageQueueService.acquireQueueLock(accountId)
+        if (!queueResult.acquired && !queueResult.skipped) {
+          // 区分 Redis 后端错误和队列超时
+          const isBackendError = queueResult.error === 'queue_backend_error'
+          const errorCode = isBackendError ? 'QUEUE_BACKEND_ERROR' : 'QUEUE_TIMEOUT'
+          const errorType = isBackendError ? 'queue_backend_error' : 'queue_timeout'
+          const errorMessage = isBackendError
+            ? 'Queue service temporarily unavailable, please retry later'
+            : 'User message queue wait timeout, please retry later'
+          const statusCode = isBackendError ? 500 : 503
+
+          // 结构化性能日志，用于后续统计
+          logger.performance('user_message_queue_error', {
+            errorType,
+            errorCode,
+            accountId,
+            statusCode,
+            apiKeyName: apiKeyData.name,
+            backendError: isBackendError ? queueResult.errorMessage : undefined
+          })
+
+          logger.warn(
+            `📬 User message queue ${errorType} for console account ${accountId}, key: ${apiKeyData.name}`,
+            isBackendError ? { backendError: queueResult.errorMessage } : {}
+          )
+          return {
+            statusCode,
+            headers: {
+              'Content-Type': 'application/json',
+              'x-user-message-queue-error': errorType
+            },
+            body: JSON.stringify({
+              type: 'error',
+              error: {
+                type: errorType,
+                code: errorCode,
+                message: errorMessage
+              }
+            }),
+            accountId
+          }
+        }
+        if (queueResult.acquired && !queueResult.skipped) {
+          queueLockAcquired = true
+          queueRequestId = queueResult.requestId
+          logger.debug(
+            `📬 User message queue lock acquired for console account ${accountId}, requestId: ${queueRequestId}`
+          )
+        }
+      }
+
       // 获取账户信息
       account = await claudeConsoleAccountService.getAccount(accountId)
       if (!account) {
         throw new Error('Claude Console Claude account not found')
       }
 
+      const autoProtectionDisabled = account.disableAutoProtection === true
+
       logger.info(
         `📤 Processing Claude Console API request for key: ${apiKeyData.name || apiKeyData.id}, account: ${account.name} (${accountId}), request: ${requestId}`
       )
@@ -201,6 +266,23 @@ class ClaudeConsoleRelayService {
       )
       const response = await axios(requestConfig)
 
+      // 📬 请求已发送成功，立即释放队列锁（无需等待响应处理完成）
+      // 因为 Claude API 限流基于请求发送时刻计算（RPM），不是请求完成时刻
+      if (queueLockAcquired && queueRequestId && accountId) {
+        try {
+          await userMessageQueueService.releaseQueueLock(accountId, queueRequestId)
+          queueLockAcquired = false // 标记已释放，防止 finally 重复释放
+          logger.debug(
+            `📬 User message queue lock released early for console account ${accountId}, requestId: ${queueRequestId}`
+          )
+        } catch (releaseError) {
+          logger.error(
+            `❌ Failed to release user message queue lock early for console account ${accountId}:`,
+            releaseError.message
+          )
+        }
+      }
+
       // 移除监听器（请求成功完成）
       if (clientRequest) {
         clientRequest.removeListener('close', handleClientDisconnect)
@@ -248,27 +330,41 @@ class ClaudeConsoleRelayService {
 
       // 检查错误状态并相应处理
       if (response.status === 401) {
-        logger.warn(`🚫 Unauthorized error detected for Claude Console account ${accountId}`)
-        await claudeConsoleAccountService.markAccountUnauthorized(accountId)
+        logger.warn(
+          `🚫 Unauthorized error detected for Claude Console account ${accountId}${autoProtectionDisabled ? ' (auto-protection disabled, skipping status change)' : ''}`
+        )
+        if (!autoProtectionDisabled) {
+          await claudeConsoleAccountService.markAccountUnauthorized(accountId)
+        }
       } else if (accountDisabledError) {
         logger.error(
-          `🚫 Account disabled error (400) detected for Claude Console account ${accountId}, marking as blocked`
+          `🚫 Account disabled error (400) detected for Claude Console account ${accountId}${autoProtectionDisabled ? ' (auto-protection disabled, skipping status change)' : ''}`
         )
         // 传入完整的错误详情到 webhook
         const errorDetails =
           typeof response.data === 'string' ? response.data : JSON.stringify(response.data)
-        await claudeConsoleAccountService.markConsoleAccountBlocked(accountId, errorDetails)
+        if (!autoProtectionDisabled) {
+          await claudeConsoleAccountService.markConsoleAccountBlocked(accountId, errorDetails)
+        }
       } else if (response.status === 429) {
-        logger.warn(`🚫 Rate limit detected for Claude Console account ${accountId}`)
+        logger.warn(
+          `🚫 Rate limit detected for Claude Console account ${accountId}${autoProtectionDisabled ? ' (auto-protection disabled, skipping status change)' : ''}`
+        )
         // 收到429先检查是否因为超过了手动配置的每日额度
         await claudeConsoleAccountService.checkQuotaUsage(accountId).catch((err) => {
           logger.error('❌ Failed to check quota after 429 error:', err)
         })
 
-        await claudeConsoleAccountService.markAccountRateLimited(accountId)
+        if (!autoProtectionDisabled) {
+          await claudeConsoleAccountService.markAccountRateLimited(accountId)
+        }
       } else if (response.status === 529) {
-        logger.warn(`🚫 Overload error detected for Claude Console account ${accountId}`)
-        await claudeConsoleAccountService.markAccountOverloaded(accountId)
+        logger.warn(
+          `🚫 Overload error detected for Claude Console account ${accountId}${autoProtectionDisabled ? ' (auto-protection disabled, skipping status change)' : ''}`
+        )
+        if (!autoProtectionDisabled) {
+          await claudeConsoleAccountService.markAccountOverloaded(accountId)
+        }
       } else if (response.status === 200 || response.status === 201) {
         // 如果请求成功，检查并移除错误状态
         const isRateLimited = await claudeConsoleAccountService.isAccountRateLimited(accountId)
@@ -350,6 +446,21 @@ class ClaudeConsoleRelayService {
           )
         }
       }
+
+      // 📬 释放用户消息队列锁（兜底，正常情况下已在请求发送后提前释放）
+      if (queueLockAcquired && queueRequestId && accountId) {
+        try {
+          await userMessageQueueService.releaseQueueLock(accountId, queueRequestId)
+          logger.debug(
+            `📬 User message queue lock released in finally for console account ${accountId}, requestId: ${queueRequestId}`
+          )
+        } catch (releaseError) {
+          logger.error(
+            `❌ Failed to release user message queue lock for account ${accountId}:`,
+            releaseError.message
+          )
+        }
+      }
     }
   }
 
@@ -368,8 +479,71 @@ class ClaudeConsoleRelayService {
     const requestId = uuidv4() // 用于并发追踪
     let concurrencyAcquired = false
     let leaseRefreshInterval = null // 租约刷新定时器
+    let queueLockAcquired = false
+    let queueRequestId = null
 
     try {
+      // 📬 用户消息队列处理：如果是用户消息请求，需要获取队列锁
+      if (userMessageQueueService.isUserMessageRequest(requestBody)) {
+        // 校验 accountId 非空，避免空值污染队列锁键
+        if (!accountId || accountId === '') {
+          logger.error(
+            '❌ accountId missing for queue lock in console relayStreamRequestWithUsageCapture'
+          )
+          throw new Error('accountId missing for queue lock')
+        }
+        const queueResult = await userMessageQueueService.acquireQueueLock(accountId)
+        if (!queueResult.acquired && !queueResult.skipped) {
+          // 区分 Redis 后端错误和队列超时
+          const isBackendError = queueResult.error === 'queue_backend_error'
+          const errorCode = isBackendError ? 'QUEUE_BACKEND_ERROR' : 'QUEUE_TIMEOUT'
+          const errorType = isBackendError ? 'queue_backend_error' : 'queue_timeout'
+          const errorMessage = isBackendError
+            ? 'Queue service temporarily unavailable, please retry later'
+            : 'User message queue wait timeout, please retry later'
+          const statusCode = isBackendError ? 500 : 503
+
+          // 结构化性能日志，用于后续统计
+          logger.performance('user_message_queue_error', {
+            errorType,
+            errorCode,
+            accountId,
+            statusCode,
+            stream: true,
+            apiKeyName: apiKeyData.name,
+            backendError: isBackendError ? queueResult.errorMessage : undefined
+          })
+
+          logger.warn(
+            `📬 User message queue ${errorType} for console account ${accountId} (stream), key: ${apiKeyData.name}`,
+            isBackendError ? { backendError: queueResult.errorMessage } : {}
+          )
+          if (!responseStream.headersSent) {
+            const existingConnection = responseStream.getHeader
+              ? responseStream.getHeader('Connection')
+              : null
+            responseStream.writeHead(statusCode, {
+              'Content-Type': 'text/event-stream',
+              'Cache-Control': 'no-cache',
+              Connection: existingConnection || 'keep-alive',
+              'x-user-message-queue-error': errorType
+            })
+          }
+          const errorEvent = `event: error\ndata: ${JSON.stringify({ type: 'error', error: { type: errorType, code: errorCode, message: errorMessage } })}\n\n`
+          responseStream.write(errorEvent)
+          responseStream.write('data: [DONE]\n\n')
+          responseStream.end()
+          return
+        }
+        if (queueResult.acquired && !queueResult.skipped) {
+          queueLockAcquired = true
+          queueRequestId = queueResult.requestId
+          logger.debug(
+            `📬 User message queue lock acquired for console account ${accountId} (stream), requestId: ${queueRequestId}`
+          )
+        }
+      }
+
       // 获取账户信息
       account = await claudeConsoleAccountService.getAccount(accountId)
       if (!account) {
@@ -467,16 +641,40 @@ class ClaudeConsoleRelayService {
         accountId,
         usageCallback,
         streamTransformer,
-        options
+        options,
+        // 📬 回调：在收到响应头时释放队列锁
+        async () => {
+          if (queueLockAcquired && queueRequestId && accountId) {
+            try {
+              await userMessageQueueService.releaseQueueLock(accountId, queueRequestId)
+              queueLockAcquired = false // 标记已释放，防止 finally 重复释放
+              logger.debug(
+                `📬 User message queue lock released early for console stream account ${accountId}, requestId: ${queueRequestId}`
+              )
+            } catch (releaseError) {
+              logger.error(
+                `❌ Failed to release user message queue lock early for console stream account ${accountId}:`,
+                releaseError.message
+              )
+            }
+          }
+        }
       )
 
       // 更新最后使用时间
       await this._updateLastUsedTime(accountId)
     } catch (error) {
-      logger.error(
-        `❌ Claude Console stream relay failed (Account: ${account?.name || accountId}):`,
-        error
-      )
+      // 客户端主动断开连接是正常情况，使用 INFO 级别
+      if (error.message === 'Client disconnected') {
+        logger.info(
+          `🔌 Claude Console stream relay ended: Client disconnected (Account: ${account?.name || accountId})`
+        )
+      } else {
+        logger.error(
+          `❌ Claude Console stream relay failed (Account: ${account?.name || accountId}):`,
+          error
+        )
+      }
       throw error
     } finally {
       // 🛑 清理租约刷新定时器
@@ -501,6 +699,21 @@ class ClaudeConsoleRelayService {
           )
         }
       }
+
+      // 📬 释放用户消息队列锁（兜底，正常情况下已在收到响应头后提前释放）
+      if (queueLockAcquired && queueRequestId && accountId) {
+        try {
+          await userMessageQueueService.releaseQueueLock(accountId, queueRequestId)
+          logger.debug(
+            `📬 User message queue lock released in finally for console stream account ${accountId}, requestId: ${queueRequestId}`
+          )
+        } catch (releaseError) {
+          logger.error(
+            `❌ Failed to release user message queue lock for stream account ${accountId}:`,
+            releaseError.message
+          )
+        }
+      }
     }
   }
 
@@ -514,7 +727,8 @@ class ClaudeConsoleRelayService {
     accountId,
     usageCallback,
     streamTransformer = null,
-    requestOptions = {}
+    requestOptions = {},
+    onResponseHeaderReceived = null
   ) {
     return new Promise((resolve, reject) => {
       let aborted = false
@@ -577,8 +791,11 @@ class ClaudeConsoleRelayService {
       // 发送请求
       const request = axios(requestConfig)
 
+      // 注意：使用 .then(async ...) 模式处理响应
+      // - 内部的 releaseQueueLock 有独立的 try-catch，不会导致未捕获异常
+      // - queueLockAcquired = false 的赋值会在 finally 执行前完成（JS 单线程保证）
       request
-        .then((response) => {
+        .then(async (response) => {
           logger.debug(`🌊 Claude Console Claude stream response status: ${response.status}`)
 
           // 错误响应处理
@@ -597,6 +814,7 @@ class ClaudeConsoleRelayService {
             })
 
             response.data.on('end', async () => {
+              const autoProtectionDisabled = account.disableAutoProtection === true
               // 记录原始错误消息到日志（方便调试，包含供应商信息）
               logger.error(
                 `📝 [Stream] Upstream error response from ${account?.name || accountId}: ${errorDataForCheck.substring(0, 500)}`
@@ -609,24 +827,41 @@ class ClaudeConsoleRelayService {
               )
 
               if (response.status === 401) {
-                await claudeConsoleAccountService.markAccountUnauthorized(accountId)
+                logger.warn(
+                  `🚫 [Stream] Unauthorized error detected for Claude Console account ${accountId}${autoProtectionDisabled ? ' (auto-protection disabled, skipping status change)' : ''}`
+                )
+                if (!autoProtectionDisabled) {
+                  await claudeConsoleAccountService.markAccountUnauthorized(accountId)
+                }
               } else if (accountDisabledError) {
                 logger.error(
-                  `🚫 [Stream] Account disabled error (400) detected for Claude Console account ${accountId}, marking as blocked`
+                  `🚫 [Stream] Account disabled error (400) detected for Claude Console account ${accountId}${autoProtectionDisabled ? ' (auto-protection disabled, skipping status change)' : ''}`
                 )
                 // 传入完整的错误详情到 webhook
-                await claudeConsoleAccountService.markConsoleAccountBlocked(
-                  accountId,
-                  errorDataForCheck
-                )
+                if (!autoProtectionDisabled) {
+                  await claudeConsoleAccountService.markConsoleAccountBlocked(
+                    accountId,
+                    errorDataForCheck
+                  )
+                }
               } else if (response.status === 429) {
-                await claudeConsoleAccountService.markAccountRateLimited(accountId)
+                logger.warn(
+                  `🚫 [Stream] Rate limit detected for Claude Console account ${accountId}${autoProtectionDisabled ? ' (auto-protection disabled, skipping status change)' : ''}`
+                )
                 // 检查是否因为超过每日额度
                 claudeConsoleAccountService.checkQuotaUsage(accountId).catch((err) => {
                   logger.error('❌ Failed to check quota after 429 error:', err)
                 })
+                if (!autoProtectionDisabled) {
+                  await claudeConsoleAccountService.markAccountRateLimited(accountId)
+                }
               } else if (response.status === 529) {
-                await claudeConsoleAccountService.markAccountOverloaded(accountId)
+                logger.warn(
+                  `🚫 [Stream] Overload error detected for Claude Console account ${accountId}${autoProtectionDisabled ? ' (auto-protection disabled, skipping status change)' : ''}`
+                )
+                if (!autoProtectionDisabled) {
+                  await claudeConsoleAccountService.markAccountOverloaded(accountId)
+                }
               }
 
               // 设置响应头
@@ -648,7 +883,7 @@ class ClaudeConsoleRelayService {
                   `🧹 [Stream] [SANITIZED] Error response to client: ${JSON.stringify(sanitizedError)}`
                 )
 
-                if (!responseStream.destroyed) {
+                if (isStreamWritable(responseStream)) {
                   responseStream.write(JSON.stringify(sanitizedError))
                   responseStream.end()
                 }
@@ -656,7 +891,7 @@ class ClaudeConsoleRelayService {
                 const sanitizedText = sanitizeErrorMessage(errorDataForCheck)
                 logger.error(`🧹 [Stream] [SANITIZED] Error response to client: ${sanitizedText}`)
 
-                if (!responseStream.destroyed) {
+                if (isStreamWritable(responseStream)) {
                   responseStream.write(sanitizedText)
                   responseStream.end()
                 }
@@ -667,6 +902,19 @@ class ClaudeConsoleRelayService {
             return
           }
 
+          // 📬 收到成功响应头（HTTP 200），调用回调释放队列锁
+          // 此时请求已被 Claude API 接受并计入 RPM 配额，无需等待响应完成
+          if (onResponseHeaderReceived && typeof onResponseHeaderReceived === 'function') {
+            try {
+              await onResponseHeaderReceived()
+            } catch (callbackError) {
+              logger.error(
+                `❌ Failed to execute onResponseHeaderReceived callback for console stream account ${accountId}:`,
+                callbackError.message
+              )
+            }
+          }
+
           // 成功响应，检查并移除错误状态
           claudeConsoleAccountService.isAccountRateLimited(accountId).then((isRateLimited) => {
             if (isRateLimited) {
@@ -680,11 +928,22 @@ class ClaudeConsoleRelayService {
           })
 
           // 设置响应头
+          // ⚠️ 关键修复：尊重 auth.js 提前设置的 Connection: close
+          // 当并发队列功能启用时，auth.js 会设置 Connection: close 来禁用 Keep-Alive
           if (!responseStream.headersSent) {
+            const existingConnection = responseStream.getHeader
+              ? responseStream.getHeader('Connection')
+              : null
+            const connectionHeader = existingConnection || 'keep-alive'
+            if (existingConnection) {
+              logger.debug(
+                `🔌 [Console Stream] Preserving existing Connection header: ${existingConnection}`
+              )
+            }
             responseStream.writeHead(200, {
               'Content-Type': 'text/event-stream',
               'Cache-Control': 'no-cache',
-              Connection: 'keep-alive',
+              Connection: connectionHeader,
               'X-Accel-Buffering': 'no'
             })
           }
@@ -710,20 +969,33 @@ class ClaudeConsoleRelayService {
               buffer = lines.pop() || ''
 
               // 转发数据并解析usage
-              if (lines.length > 0 && !responseStream.destroyed) {
-                const linesToForward = lines.join('\n') + (lines.length > 0 ? '\n' : '')
+              if (lines.length > 0) {
+                // 检查流是否可写（客户端连接是否有效）
+                if (isStreamWritable(responseStream)) {
+                  const linesToForward = lines.join('\n') + (lines.length > 0 ? '\n' : '')
 
-                // 应用流转换器如果有
-                if (streamTransformer) {
-                  const transformed = streamTransformer(linesToForward)
-                  if (transformed) {
-                    responseStream.write(transformed)
+                  // 应用流转换器如果有
+                  let dataToWrite = linesToForward
+                  if (streamTransformer) {
+                    const transformed = streamTransformer(linesToForward)
+                    if (transformed) {
+                      dataToWrite = transformed
+                    } else {
+                      dataToWrite = null
+                    }
+                  }
+
+                  if (dataToWrite) {
+                    responseStream.write(dataToWrite)
                   }
                 } else {
-                  responseStream.write(linesToForward)
+                  // 客户端连接已断开，记录警告（但仍继续解析usage）
+                  logger.warn(
+                    `⚠️ [Console] Client disconnected during stream, skipping ${lines.length} lines for account: ${account?.name || accountId}`
+                  )
                 }
 
-                // 解析SSE数据寻找usage信息
+                // 解析SSE数据寻找usage信息（无论连接状态如何）
                 for (const line of lines) {
                   if (line.startsWith('data:')) {
                     const jsonStr = line.slice(5).trimStart()
@@ -831,7 +1103,7 @@ class ClaudeConsoleRelayService {
                 `❌ Error processing Claude Console stream data (Account: ${account?.name || accountId}):`,
                 error
               )
-              if (!responseStream.destroyed) {
+              if (isStreamWritable(responseStream)) {
                 // 如果有 streamTransformer（如测试请求），使用前端期望的格式
                 if (streamTransformer) {
                   responseStream.write(
@@ -854,7 +1126,7 @@ class ClaudeConsoleRelayService {
           response.data.on('end', () => {
             try {
               // 处理缓冲区中剩余的数据
-              if (buffer.trim() && !responseStream.destroyed) {
+              if (buffer.trim() && isStreamWritable(responseStream)) {
                 if (streamTransformer) {
                   const transformed = streamTransformer(buffer)
                   if (transformed) {
@@ -903,12 +1175,33 @@ class ClaudeConsoleRelayService {
               }
 
               // 确保流正确结束
-              if (!responseStream.destroyed) {
-                responseStream.end()
-              }
+              if (isStreamWritable(responseStream)) {
+                // 📊 诊断日志：流结束前状态
+                logger.info(
+                  `📤 [STREAM] Ending response | destroyed: ${responseStream.destroyed}, ` +
+                    `socketDestroyed: ${responseStream.socket?.destroyed}, ` +
+                    `socketBytesWritten: ${responseStream.socket?.bytesWritten || 0}`
+                )
 
-              logger.debug('🌊 Claude Console Claude stream response completed')
-              resolve()
+                // 禁用 Nagle 算法确保数据立即发送
+                if (responseStream.socket && !responseStream.socket.destroyed) {
+                  responseStream.socket.setNoDelay(true)
+                }
+
+                // 等待数据完全 flush 到客户端后再 resolve
+                responseStream.end(() => {
+                  logger.info(
+                    `✅ [STREAM] Response ended and flushed | socketBytesWritten: ${responseStream.socket?.bytesWritten || 'unknown'}`
+                  )
+                  resolve()
+                })
+              } else {
+                // 连接已断开，记录警告
+                logger.warn(
+                  `⚠️ [Console] Client disconnected before stream end, data may not have been received | account: ${account?.name || accountId}`
+                )
+                resolve()
+              }
             } catch (error) {
               logger.error('❌ Error processing stream end:', error)
               reject(error)
@@ -920,7 +1213,7 @@ class ClaudeConsoleRelayService {
               `❌ Claude Console stream error (Account: ${account?.name || accountId}):`,
               error
             )
-            if (!responseStream.destroyed) {
+            if (isStreamWritable(responseStream)) {
               // 如果有 streamTransformer（如测试请求），使用前端期望的格式
               if (streamTransformer) {
                 responseStream.write(
@@ -968,14 +1261,17 @@ class ClaudeConsoleRelayService {
 
           // 发送错误响应
           if (!responseStream.headersSent) {
+            const existingConnection = responseStream.getHeader
+              ? responseStream.getHeader('Connection')
+              : null
             responseStream.writeHead(error.response?.status || 500, {
               'Content-Type': 'text/event-stream',
               'Cache-Control': 'no-cache',
-              Connection: 'keep-alive'
+              Connection: existingConnection || 'keep-alive'
             })
           }
 
-          if (!responseStream.destroyed) {
+          if (isStreamWritable(responseStream)) {
             // 如果有 streamTransformer（如测试请求），使用前端期望的格式
             if (streamTransformer) {
               responseStream.write(
@@ -1007,30 +1303,9 @@ class ClaudeConsoleRelayService {
 
   // 🔧 过滤客户端请求头
   _filterClientHeaders(clientHeaders) {
-    const sensitiveHeaders = [
-      'content-type',
-      'user-agent',
-      'authorization',
-      'x-api-key',
-      'host',
-      'content-length',
-      'connection',
-      'proxy-authorization',
-      'content-encoding',
-      'transfer-encoding',
-      'anthropic-version'
-    ]
-
-    const filteredHeaders = {}
-
-    Object.keys(clientHeaders || {}).forEach((key) => {
-      const lowerKey = key.toLowerCase()
-      if (!sensitiveHeaders.includes(lowerKey)) {
-        filteredHeaders[key] = clientHeaders[key]
-      }
-    })
-
-    return filteredHeaders
+    // 使用统一的 headerFilter 工具类（白名单模式）
+    // 与 claudeRelayService 保持一致，避免透传 CDN headers 触发上游 API 安全检查
+    return filterForClaude(clientHeaders)
   }
 
   // 🕐 更新最后使用时间
@@ -1113,55 +1388,11 @@ class ClaudeConsoleRelayService {
     }
   }
 
-  // 🧪 测试账号连接（供Admin API使用，独立处理以确保错误时也返回SSE格式）
+  // 🧪 测试账号连接（供Admin API使用）
   async testAccountConnection(accountId, responseStream) {
-    const testRequestBody = {
-      model: 'claude-sonnet-4-5-20250929',
-      max_tokens: 32000,
-      stream: true,
-      messages: [
-        {
-          role: 'user',
-          content: 'hi'
-        }
-      ],
-      system: [
-        {
-          type: 'text',
-          text: "You are Claude Code, Anthropic's official CLI for Claude."
-        }
-      ]
-    }
-
-    // 辅助函数：发送 SSE 事件
-    const sendSSEEvent = (type, data) => {
-      if (!responseStream.destroyed && !responseStream.writableEnded) {
-        try {
-          responseStream.write(`data: ${JSON.stringify({ type, ...data })}\n\n`)
-        } catch {
-          // 忽略写入错误
-        }
-      }
-    }
-
-    // 辅助函数：结束测试并关闭流
-    const endTest = (success, error = null) => {
-      if (!responseStream.destroyed && !responseStream.writableEnded) {
-        try {
-          if (success) {
-            sendSSEEvent('test_complete', { success: true })
-          } else {
-            sendSSEEvent('test_complete', { success: false, error: error || '测试失败' })
-          }
-          responseStream.end()
-        } catch {
-          // 忽略写入错误
-        }
-      }
-    }
+    const { sendStreamTestRequest } = require('../utils/testPayloadHelper')
 
     try {
-      // 获取账户信息
       const account = await claudeConsoleAccountService.getAccount(accountId)
       if (!account) {
         throw new Error('Account not found')
@@ -1169,178 +1400,32 @@ class ClaudeConsoleRelayService {
 
       logger.info(`🧪 Testing Claude Console account connection: ${account.name} (${accountId})`)
 
-      // 创建代理agent
-      const proxyAgent = claudeConsoleAccountService._createProxyAgent(account.proxy)
-
-      // 设置响应头
-      if (!responseStream.headersSent) {
-        responseStream.writeHead(200, {
-          'Content-Type': 'text/event-stream',
-          'Cache-Control': 'no-cache',
-          Connection: 'keep-alive',
-          'X-Accel-Buffering': 'no'
-        })
-      }
-
-      // 发送测试开始事件
-      sendSSEEvent('test_start', {})
-
-      // 构建完整的API URL
       const cleanUrl = account.apiUrl.replace(/\/$/, '')
-      const apiEndpoint = cleanUrl.endsWith('/v1/messages') ? cleanUrl : `${cleanUrl}/v1/messages`
+      const apiUrl = cleanUrl.endsWith('/v1/messages')
+        ? cleanUrl
+        : `${cleanUrl}/v1/messages?beta=true`
 
-      // 决定使用的 User-Agent
-      const userAgent = account.userAgent || this.defaultUserAgent
-
-      // 准备请求配置
-      const requestConfig = {
-        method: 'POST',
-        url: apiEndpoint,
-        data: testRequestBody,
-        headers: {
-          'Content-Type': 'application/json',
-          'anthropic-version': '2023-06-01',
-          'User-Agent': userAgent
-        },
-        timeout: 30000, // 测试请求使用较短超时
-        responseType: 'stream',
-        validateStatus: () => true
-      }
-
-      if (proxyAgent) {
-        requestConfig.httpAgent = proxyAgent
-        requestConfig.httpsAgent = proxyAgent
-        requestConfig.proxy = false
-      }
-
-      // 设置认证方式
-      if (account.apiKey && account.apiKey.startsWith('sk-ant-')) {
-        requestConfig.headers['x-api-key'] = account.apiKey
-      } else {
-        requestConfig.headers['Authorization'] = `Bearer ${account.apiKey}`
-      }
-
-      // 发送请求
-      const response = await axios(requestConfig)
-
-      logger.debug(`🌊 Claude Console test response status: ${response.status}`)
-
-      // 处理非200响应
-      if (response.status !== 200) {
-        logger.error(
-          `❌ Claude Console API returned error status: ${response.status} | Account: ${account?.name || accountId}`
-        )
-
-        // 收集错误响应数据
-        return new Promise((resolve) => {
-          const errorChunks = []
-
-          response.data.on('data', (chunk) => {
-            errorChunks.push(chunk)
-          })
-
-          response.data.on('end', () => {
-            try {
-              const fullErrorData = Buffer.concat(errorChunks).toString()
-              logger.error(
-                `📝 [Test] Upstream error response from ${account?.name || accountId}: ${fullErrorData.substring(0, 500)}`
-              )
-
-              // 尝试解析错误信息
-              let errorMessage = `API Error: ${response.status}`
-              try {
-                const errorJson = JSON.parse(fullErrorData)
-                // 直接提取所有可能的错误信息字段
-                errorMessage =
-                  errorJson.message ||
-                  errorJson.error?.message ||
-                  errorJson.statusMessage ||
-                  errorJson.error ||
-                  (typeof errorJson === 'string' ? errorJson : JSON.stringify(errorJson))
-              } catch {
-                errorMessage = fullErrorData.substring(0, 200) || `API Error: ${response.status}`
-              }
-
-              endTest(false, errorMessage)
-              resolve()
-            } catch {
-              endTest(false, `API Error: ${response.status}`)
-              resolve()
-            }
-          })
-
-          response.data.on('error', (err) => {
-            endTest(false, err.message || '流读取错误')
-            resolve()
-          })
-        })
-      }
-
-      // 处理成功的流式响应
-      return new Promise((resolve) => {
-        let buffer = ''
-
-        response.data.on('data', (chunk) => {
-          try {
-            buffer += chunk.toString()
-            const lines = buffer.split('\n')
-            buffer = lines.pop() || ''
-
-            for (const line of lines) {
-              if (!line.startsWith('data: ')) {
-                continue
-              }
-
-              const jsonStr = line.substring(6).trim()
-              if (!jsonStr || jsonStr === '[DONE]') {
-                continue
-              }
-
-              try {
-                const data = JSON.parse(jsonStr)
-
-                // 转换 content_block_delta 为 content
-                if (data.type === 'content_block_delta' && data.delta && data.delta.text) {
-                  sendSSEEvent('content', { text: data.delta.text })
-                }
-
-                // 处理消息完成
-                if (data.type === 'message_stop') {
-                  endTest(true)
-                }
-
-                // 处理错误事件
-                if (data.type === 'error') {
-                  const errorMsg = data.error?.message || data.message || '未知错误'
-                  endTest(false, errorMsg)
-                }
-              } catch {
-                // 忽略解析错误
-              }
-            }
-          } catch {
-            // 忽略处理错误
-          }
-        })
-
-        response.data.on('end', () => {
-          logger.info(`✅ Test request completed for account: ${account.name}`)
-          // 如果还没结束，发送完成事件
-          if (!responseStream.destroyed && !responseStream.writableEnded) {
-            endTest(true)
-          }
-          resolve()
-        })
-
-        response.data.on('error', (err) => {
-          logger.error(`❌ Test stream error:`, err)
-          endTest(false, err.message || '流处理错误')
-          resolve()
-        })
+      await sendStreamTestRequest({
+        apiUrl,
+        authorization: `Bearer ${account.apiKey}`,
+        responseStream,
+        proxyAgent: claudeConsoleAccountService._createProxyAgent(account.proxy),
+        extraHeaders: account.userAgent ? { 'User-Agent': account.userAgent } : {}
       })
     } catch (error) {
       logger.error(`❌ Test account connection failed:`, error)
-      endTest(false, error.message || '测试失败')
+      if (!responseStream.headersSent) {
+        responseStream.writeHead(200, {
+          'Content-Type': 'text/event-stream',
+          'Cache-Control': 'no-cache'
+        })
+      }
+      if (isStreamWritable(responseStream)) {
+        responseStream.write(
+          `data: ${JSON.stringify({ type: 'test_complete', success: false, error: error.message })}\n\n`
+        )
+        responseStream.end()
+      }
     }
   }
 

src/services/claudeRelayConfigService.js

@@ -0,0 +1,453 @@
+/**
+ * Claude 转发配置服务
+ * 管理全局 Claude Code 限制和会话绑定配置
+ */
+
+const redis = require('../models/redis')
+const logger = require('../utils/logger')
+
+const CONFIG_KEY = 'claude_relay_config'
+const SESSION_BINDING_PREFIX = 'original_session_binding:'
+
+// 默认配置
+const DEFAULT_CONFIG = {
+  claudeCodeOnlyEnabled: false,
+  globalSessionBindingEnabled: false,
+  sessionBindingErrorMessage: '你的本地session已污染，请清理后使用。',
+  sessionBindingTtlDays: 30, // 会话绑定 TTL（天），默认30天
+  // 用户消息队列配置
+  userMessageQueueEnabled: false, // 是否启用用户消息队列（默认关闭）
+  userMessageQueueDelayMs: 200, // 请求间隔（毫秒）
+  userMessageQueueTimeoutMs: 5000, // 队列等待超时（毫秒），优化后锁持有时间短无需长等待
+  userMessageQueueLockTtlMs: 5000, // 锁TTL（毫秒），请求发送后立即释放无需长TTL
+  // 并发请求排队配置
+  concurrentRequestQueueEnabled: false, // 是否启用并发请求排队（默认关闭）
+  concurrentRequestQueueMaxSize: 3, // 固定最小排队数（默认3）
+  concurrentRequestQueueMaxSizeMultiplier: 0, // 并发数的倍数（默认0，仅使用固定值）
+  concurrentRequestQueueTimeoutMs: 10000, // 排队超时（毫秒，默认10秒）
+  concurrentRequestQueueMaxRedisFailCount: 5, // 连续 Redis 失败阈值（默认5次）
+  // 排队健康检查配置
+  concurrentRequestQueueHealthCheckEnabled: true, // 是否启用排队健康检查（默认开启）
+  concurrentRequestQueueHealthThreshold: 0.8, // 健康检查阈值（P90 >= 超时 × 阈值时拒绝新请求）
+  updatedAt: null,
+  updatedBy: null
+}
+
+// 内存缓存（避免频繁 Redis 查询）
+let configCache = null
+let configCacheTime = 0
+const CONFIG_CACHE_TTL = 60000 // 1分钟缓存
+
+class ClaudeRelayConfigService {
+  /**
+   * 从 metadata.user_id 中提取原始 sessionId
+   * 格式: user_{64位十六进制}_account__session_{uuid}
+   * @param {Object} requestBody - 请求体
+   * @returns {string|null} 原始 sessionId 或 null
+   */
+  extractOriginalSessionId(requestBody) {
+    if (!requestBody?.metadata?.user_id) {
+      return null
+    }
+
+    const userId = requestBody.metadata.user_id
+    const match = userId.match(/session_([a-f0-9-]{36})$/i)
+    return match ? match[1] : null
+  }
+
+  /**
+   * 获取配置（带缓存）
+   * @returns {Promise<Object>} 配置对象
+   */
+  async getConfig() {
+    try {
+      // 检查缓存
+      if (configCache && Date.now() - configCacheTime < CONFIG_CACHE_TTL) {
+        return configCache
+      }
+
+      const client = redis.getClient()
+      if (!client) {
+        logger.warn('⚠️ Redis not connected, using default config')
+        return { ...DEFAULT_CONFIG }
+      }
+
+      const data = await client.get(CONFIG_KEY)
+
+      if (data) {
+        configCache = { ...DEFAULT_CONFIG, ...JSON.parse(data) }
+      } else {
+        configCache = { ...DEFAULT_CONFIG }
+      }
+
+      configCacheTime = Date.now()
+      return configCache
+    } catch (error) {
+      logger.error('❌ Failed to get Claude relay config:', error)
+      return { ...DEFAULT_CONFIG }
+    }
+  }
+
+  /**
+   * 更新配置
+   * @param {Object} newConfig - 新配置
+   * @param {string} updatedBy - 更新者
+   * @returns {Promise<Object>} 更新后的配置
+   */
+  async updateConfig(newConfig, updatedBy) {
+    try {
+      const client = redis.getClientSafe()
+      const currentConfig = await this.getConfig()
+
+      const updatedConfig = {
+        ...currentConfig,
+        ...newConfig,
+        updatedAt: new Date().toISOString(),
+        updatedBy
+      }
+
+      await client.set(CONFIG_KEY, JSON.stringify(updatedConfig))
+
+      // 更新缓存
+      configCache = updatedConfig
+      configCacheTime = Date.now()
+
+      logger.info(`✅ Claude relay config updated by ${updatedBy}:`, {
+        claudeCodeOnlyEnabled: updatedConfig.claudeCodeOnlyEnabled,
+        globalSessionBindingEnabled: updatedConfig.globalSessionBindingEnabled,
+        concurrentRequestQueueEnabled: updatedConfig.concurrentRequestQueueEnabled
+      })
+
+      return updatedConfig
+    } catch (error) {
+      logger.error('❌ Failed to update Claude relay config:', error)
+      throw error
+    }
+  }
+
+  /**
+   * 检查是否启用全局 Claude Code 限制
+   * @returns {Promise<boolean>}
+   */
+  async isClaudeCodeOnlyEnabled() {
+    const cfg = await this.getConfig()
+    return cfg.claudeCodeOnlyEnabled === true
+  }
+
+  /**
+   * 检查是否启用全局会话绑定
+   * @returns {Promise<boolean>}
+   */
+  async isGlobalSessionBindingEnabled() {
+    const cfg = await this.getConfig()
+    return cfg.globalSessionBindingEnabled === true
+  }
+
+  /**
+   * 获取会话绑定错误信息
+   * @returns {Promise<string>}
+   */
+  async getSessionBindingErrorMessage() {
+    const cfg = await this.getConfig()
+    return cfg.sessionBindingErrorMessage || DEFAULT_CONFIG.sessionBindingErrorMessage
+  }
+
+  /**
+   * 获取原始会话绑定
+   * @param {string} originalSessionId - 原始会话ID
+   * @returns {Promise<Object|null>} 绑定信息或 null
+   */
+  async getOriginalSessionBinding(originalSessionId) {
+    if (!originalSessionId) {
+      return null
+    }
+
+    try {
+      const client = redis.getClient()
+      if (!client) {
+        return null
+      }
+
+      const key = `${SESSION_BINDING_PREFIX}${originalSessionId}`
+      const data = await client.get(key)
+
+      if (data) {
+        return JSON.parse(data)
+      }
+      return null
+    } catch (error) {
+      logger.error(`❌ Failed to get session binding for ${originalSessionId}:`, error)
+      return null
+    }
+  }
+
+  /**
+   * 设置原始会话绑定
+   * @param {string} originalSessionId - 原始会话ID
+   * @param {string} accountId - 账户ID
+   * @param {string} accountType - 账户类型
+   * @returns {Promise<Object>} 绑定信息
+   */
+  async setOriginalSessionBinding(originalSessionId, accountId, accountType) {
+    if (!originalSessionId || !accountId || !accountType) {
+      throw new Error('Invalid parameters for session binding')
+    }
+
+    try {
+      const client = redis.getClientSafe()
+      const key = `${SESSION_BINDING_PREFIX}${originalSessionId}`
+
+      const binding = {
+        accountId,
+        accountType,
+        createdAt: new Date().toISOString(),
+        lastUsedAt: new Date().toISOString()
+      }
+
+      // 使用配置的 TTL（默认30天）
+      const cfg = await this.getConfig()
+      const ttlDays = cfg.sessionBindingTtlDays || DEFAULT_CONFIG.sessionBindingTtlDays
+      const ttlSeconds = Math.floor(ttlDays * 24 * 3600)
+
+      await client.set(key, JSON.stringify(binding), 'EX', ttlSeconds)
+
+      logger.info(
+        `🔗 Session binding created: ${originalSessionId} -> ${accountId} (${accountType})`
+      )
+
+      return binding
+    } catch (error) {
+      logger.error(`❌ Failed to set session binding for ${originalSessionId}:`, error)
+      throw error
+    }
+  }
+
+  /**
+   * 更新会话绑定的最后使用时间（续期）
+   * @param {string} originalSessionId - 原始会话ID
+   */
+  async touchOriginalSessionBinding(originalSessionId) {
+    if (!originalSessionId) {
+      return
+    }
+
+    try {
+      const binding = await this.getOriginalSessionBinding(originalSessionId)
+      if (!binding) {
+        return
+      }
+
+      binding.lastUsedAt = new Date().toISOString()
+
+      const client = redis.getClientSafe()
+      const key = `${SESSION_BINDING_PREFIX}${originalSessionId}`
+
+      // 使用配置的 TTL（默认30天）
+      const cfg = await this.getConfig()
+      const ttlDays = cfg.sessionBindingTtlDays || DEFAULT_CONFIG.sessionBindingTtlDays
+      const ttlSeconds = Math.floor(ttlDays * 24 * 3600)
+
+      await client.set(key, JSON.stringify(binding), 'EX', ttlSeconds)
+    } catch (error) {
+      logger.warn(`⚠️ Failed to touch session binding for ${originalSessionId}:`, error)
+    }
+  }
+
+  /**
+   * 检查原始会话是否已绑定
+   * @param {string} originalSessionId - 原始会话ID
+   * @returns {Promise<boolean>}
+   */
+  async isOriginalSessionBound(originalSessionId) {
+    const binding = await this.getOriginalSessionBinding(originalSessionId)
+    return binding !== null
+  }
+
+  /**
+   * 验证绑定的账户是否可用
+   * @param {Object} binding - 绑定信息
+   * @returns {Promise<boolean>}
+   */
+  async validateBoundAccount(binding) {
+    if (!binding || !binding.accountId || !binding.accountType) {
+      return false
+    }
+
+    try {
+      const { accountType } = binding
+      const { accountId } = binding
+
+      let accountService
+      switch (accountType) {
+        case 'claude-official':
+          accountService = require('./claudeAccountService')
+          break
+        case 'claude-console':
+          accountService = require('./claudeConsoleAccountService')
+          break
+        case 'bedrock':
+          accountService = require('./bedrockAccountService')
+          break
+        case 'ccr':
+          accountService = require('./ccrAccountService')
+          break
+        default:
+          logger.warn(`Unknown account type for validation: ${accountType}`)
+          return false
+      }
+
+      const account = await accountService.getAccount(accountId)
+
+      // getAccount() 直接返回账户数据对象或 null，不是 { success, data } 格式
+      if (!account) {
+        logger.warn(`Session binding account not found: ${accountId} (${accountType})`)
+        return false
+      }
+
+      const accountData = account
+
+      // 检查账户是否激活
+      if (accountData.isActive === false || accountData.isActive === 'false') {
+        logger.warn(
+          `Session binding account not active: ${accountId} (${accountType}), isActive: ${accountData.isActive}`
+        )
+        return false
+      }
+
+      // 检查账户状态（如果存在）
+      if (accountData.status && accountData.status === 'error') {
+        logger.warn(
+          `Session binding account has error status: ${accountId} (${accountType}), status: ${accountData.status}`
+        )
+        return false
+      }
+
+      return true
+    } catch (error) {
+      logger.error(`❌ Failed to validate bound account ${binding.accountId}:`, error)
+      return false
+    }
+  }
+
+  /**
+   * 验证新会话请求
+   * @param {Object} _requestBody - 请求体（预留参数，当前未使用）
+   * @param {string} originalSessionId - 原始会话ID
+   * @returns {Promise<Object>} { valid: boolean, error?: string, binding?: object, isNewSession?: boolean }
+   */
+  async validateNewSession(_requestBody, originalSessionId) {
+    const cfg = await this.getConfig()
+
+    if (!cfg.globalSessionBindingEnabled) {
+      return { valid: true }
+    }
+
+    // 如果没有 sessionId，跳过验证（可能是非 Claude Code 客户端）
+    if (!originalSessionId) {
+      return { valid: true }
+    }
+
+    const existingBinding = await this.getOriginalSessionBinding(originalSessionId)
+
+    // 如果会话已存在绑定
+    if (existingBinding) {
+      // ⚠️ 只有 claude-official 类型账户受全局会话绑定限制
+      // 其他类型（bedrock, ccr, claude-console等）忽略绑定，走正常调度
+      if (existingBinding.accountType !== 'claude-official') {
+        logger.info(
+          `🔗 Session binding ignored for non-official account type: ${existingBinding.accountType}`
+        )
+        return { valid: true }
+      }
+
+      const accountValid = await this.validateBoundAccount(existingBinding)
+
+      if (!accountValid) {
+        return {
+          valid: false,
+          error: cfg.sessionBindingErrorMessage,
+          code: 'SESSION_BINDING_INVALID'
+        }
+      }
+
+      // 续期
+      await this.touchOriginalSessionBinding(originalSessionId)
+
+      // 已有绑定，允许继续（这是正常的会话延续）
+      return { valid: true, binding: existingBinding }
+    }
+
+    // 没有绑定，是新会话
+    // 注意：messages.length 检查在此处无法执行，因为我们不知道最终会调度到哪种账户类型
+    // 绑定会在调度后创建，仅针对 claude-official 账户
+    return { valid: true, isNewSession: true }
+  }
+
+  /**
+   * 删除原始会话绑定
+   * @param {string} originalSessionId - 原始会话ID
+   */
+  async deleteOriginalSessionBinding(originalSessionId) {
+    if (!originalSessionId) {
+      return
+    }
+
+    try {
+      const client = redis.getClient()
+      if (!client) {
+        return
+      }
+
+      const key = `${SESSION_BINDING_PREFIX}${originalSessionId}`
+      await client.del(key)
+      logger.info(`🗑️ Session binding deleted: ${originalSessionId}`)
+    } catch (error) {
+      logger.error(`❌ Failed to delete session binding for ${originalSessionId}:`, error)
+    }
+  }
+
+  /**
+   * 获取会话绑定统计
+   * @returns {Promise<Object>}
+   */
+  async getSessionBindingStats() {
+    try {
+      const client = redis.getClient()
+      if (!client) {
+        return { totalBindings: 0 }
+      }
+
+      let cursor = '0'
+      let count = 0
+
+      do {
+        const [newCursor, keys] = await client.scan(
+          cursor,
+          'MATCH',
+          `${SESSION_BINDING_PREFIX}*`,
+          'COUNT',
+          100
+        )
+        cursor = newCursor
+        count += keys.length
+      } while (cursor !== '0')
+
+      return {
+        totalBindings: count
+      }
+    } catch (error) {
+      logger.error('❌ Failed to get session binding stats:', error)
+      return { totalBindings: 0 }
+    }
+  }
+
+  /**
+   * 清除配置缓存（用于测试或强制刷新）
+   */
+  clearCache() {
+    configCache = null
+    configCacheTime = 0
+  }
+}
+
+module.exports = new ClaudeRelayConfigService()

src/services/droidAccountService.js

@@ -556,7 +556,8 @@ class DroidAccountService {
       tokenType = 'Bearer',
       authenticationMethod = '',
       expiresIn = null,
-      apiKeys = []
+      apiKeys = [],
+      userAgent = '' // 自定义 User-Agent
     } = options
 
     const accountId = uuidv4()
@@ -832,7 +833,8 @@ class DroidAccountService {
           : '',
       apiKeys: hasApiKeys ? JSON.stringify(apiKeyEntries) : '',
       apiKeyCount: hasApiKeys ? String(apiKeyEntries.length) : '0',
-      apiKeyStrategy: hasApiKeys ? 'random_sticky' : ''
+      apiKeyStrategy: hasApiKeys ? 'random_sticky' : '',
+      userAgent: userAgent || '' // 自定义 User-Agent
     }
 
     await redis.setDroidAccount(accountId, accountData)
@@ -931,6 +933,11 @@ class DroidAccountService {
       sanitizedUpdates.endpointType = this._sanitizeEndpointType(sanitizedUpdates.endpointType)
     }
 
+    // 处理 userAgent 字段
+    if (typeof sanitizedUpdates.userAgent === 'string') {
+      sanitizedUpdates.userAgent = sanitizedUpdates.userAgent.trim()
+    }
+
     const parseProxyConfig = (value) => {
       if (!value) {
         return null

src/services/droidRelayService.js

@@ -26,7 +26,7 @@ class DroidRelayService {
       comm: '/o/v1/chat/completions'
     }
 
-    this.userAgent = 'factory-cli/0.19.12'
+    this.userAgent = 'factory-cli/0.32.1'
     this.systemPrompt = SYSTEM_PROMPT
     this.API_KEY_STICKY_PREFIX = 'droid_api_key'
   }
@@ -241,7 +241,8 @@ class DroidRelayService {
         accessToken,
         normalizedRequestBody,
         normalizedEndpoint,
-        clientHeaders
+        clientHeaders,
+        account
       )
 
       if (selectedApiKey) {
@@ -335,7 +336,12 @@ class DroidRelayService {
         )
       }
     } catch (error) {
-      logger.error(`❌ Droid relay error: ${error.message}`, error)
+      // 客户端主动断开连接是正常情况，使用 INFO 级别
+      if (error.message === 'Client disconnected') {
+        logger.info(`🔌 Droid relay ended: Client disconnected`)
+      } else {
+        logger.error(`❌ Droid relay error: ${error.message}`, error)
+      }
 
       const status = error?.response?.status
       if (status >= 400 && status < 500) {
@@ -633,7 +639,7 @@ class DroidRelayService {
       // 客户端断开连接时清理
       clientResponse.on('close', () => {
         if (req && !req.destroyed) {
-          req.destroy()
+          req.destroy(new Error('Client disconnected'))
         }
       })
 
@@ -737,6 +743,14 @@ class DroidRelayService {
                 currentUsageData.output_tokens = 0
               }
 
+              // Capture cache tokens from OpenAI format
+              currentUsageData.cache_read_input_tokens =
+                data.usage.input_tokens_details?.cached_tokens || 0
+              currentUsageData.cache_creation_input_tokens =
+                data.usage.input_tokens_details?.cache_creation_input_tokens ||
+                data.usage.cache_creation_input_tokens ||
+                0
+
               logger.debug('📊 Droid OpenAI usage:', currentUsageData)
             }
 
@@ -758,6 +772,14 @@ class DroidRelayService {
                 currentUsageData.output_tokens = 0
               }
 
+              // Capture cache tokens from OpenAI Response API format
+              currentUsageData.cache_read_input_tokens =
+                usage.input_tokens_details?.cached_tokens || 0
+              currentUsageData.cache_creation_input_tokens =
+                usage.input_tokens_details?.cache_creation_input_tokens ||
+                usage.cache_creation_input_tokens ||
+                0
+
               logger.debug('📊 Droid OpenAI response usage:', currentUsageData)
             }
           } catch (parseError) {
@@ -966,11 +988,13 @@ class DroidRelayService {
   /**
    * 构建请求头
    */
-  _buildHeaders(accessToken, requestBody, endpointType, clientHeaders = {}) {
+  _buildHeaders(accessToken, requestBody, endpointType, clientHeaders = {}, account = null) {
+    // 使用账户配置的 userAgent 或默认值
+    const userAgent = account?.userAgent || this.userAgent
     const headers = {
       'content-type': 'application/json',
       authorization: `Bearer ${accessToken}`,
-      'user-agent': this.userAgent,
+      'user-agent': userAgent,
       'x-factory-client': 'cli',
       connection: 'keep-alive'
     }
@@ -987,9 +1011,15 @@ class DroidRelayService {
       }
     }
 
-    // OpenAI 特定头
+    // OpenAI 特定头 - 根据模型动态选择 provider
     if (endpointType === 'openai') {
-      headers['x-api-provider'] = 'azure_openai'
+      const model = (requestBody?.model || '').toLowerCase()
+      // -max 模型使用 openai provider，其他使用 azure_openai
+      if (model.includes('-max')) {
+        headers['x-api-provider'] = 'openai'
+      } else {
+        headers['x-api-provider'] = 'azure_openai'
+      }
     }
 
     // Comm 端点根据模型动态设置 provider

src/services/geminiAccountService.js

@@ -16,11 +16,62 @@ const {
 } = require('../utils/tokenRefreshLogger')
 const tokenRefreshService = require('./tokenRefreshService')
 const LRUCache = require('../utils/lruCache')
+const antigravityClient = require('./antigravityClient')
 
-// Gemini CLI OAuth 配置 - 这些是公开的 Gemini CLI 凭据
-const OAUTH_CLIENT_ID = '681255809395-oo8ft2oprdrnp9e3aqf6av3hmdib135j.apps.googleusercontent.com'
-const OAUTH_CLIENT_SECRET = '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'
-const OAUTH_SCOPES = ['https://www.googleapis.com/auth/cloud-platform']
+// Gemini OAuth 配置 - 支持 Gemini CLI 与 Antigravity 两种 OAuth 应用
+const OAUTH_PROVIDER_GEMINI_CLI = 'gemini-cli'
+const OAUTH_PROVIDER_ANTIGRAVITY = 'antigravity'
+
+const OAUTH_PROVIDERS = {
+  [OAUTH_PROVIDER_GEMINI_CLI]: {
+    // Gemini CLI OAuth 配置（公开）
+    clientId:
+      process.env.GEMINI_OAUTH_CLIENT_ID ||
+      '681255809395-oo8ft2oprdrnp9e3aqf6av3hmdib135j.apps.googleusercontent.com',
+    clientSecret: process.env.GEMINI_OAUTH_CLIENT_SECRET || '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',
+    scopes: ['https://www.googleapis.com/auth/cloud-platform']
+  },
+  [OAUTH_PROVIDER_ANTIGRAVITY]: {
+    // Antigravity OAuth 配置（参考 gcli2api）
+    clientId:
+      process.env.ANTIGRAVITY_OAUTH_CLIENT_ID ||
+      '1071006060591-tmhssin2h21lcre235vtolojh4g403ep.apps.googleusercontent.com',
+    clientSecret:
+      process.env.ANTIGRAVITY_OAUTH_CLIENT_SECRET || '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',
+    scopes: [
+      'https://www.googleapis.com/auth/cloud-platform',
+      'https://www.googleapis.com/auth/userinfo.email',
+      'https://www.googleapis.com/auth/userinfo.profile',
+      'https://www.googleapis.com/auth/cclog',
+      'https://www.googleapis.com/auth/experimentsandconfigs'
+    ]
+  }
+}
+
+if (!process.env.GEMINI_OAUTH_CLIENT_SECRET) {
+  logger.warn(
+    '⚠️ GEMINI_OAUTH_CLIENT_SECRET 未设置，使用内置默认值（建议在生产环境通过环境变量覆盖）'
+  )
+}
+if (!process.env.ANTIGRAVITY_OAUTH_CLIENT_SECRET) {
+  logger.warn(
+    '⚠️ ANTIGRAVITY_OAUTH_CLIENT_SECRET 未设置，使用内置默认值（建议在生产环境通过环境变量覆盖）'
+  )
+}
+
+function normalizeOauthProvider(oauthProvider) {
+  if (!oauthProvider) {
+    return OAUTH_PROVIDER_GEMINI_CLI
+  }
+  return oauthProvider === OAUTH_PROVIDER_ANTIGRAVITY
+    ? OAUTH_PROVIDER_ANTIGRAVITY
+    : OAUTH_PROVIDER_GEMINI_CLI
+}
+
+function getOauthProviderConfig(oauthProvider) {
+  const normalized = normalizeOauthProvider(oauthProvider)
+  return OAUTH_PROVIDERS[normalized] || OAUTH_PROVIDERS[OAUTH_PROVIDER_GEMINI_CLI]
+}
 
 // 🌐 TCP Keep-Alive Agent 配置
 // 解决长时间流式请求中 NAT/防火墙空闲超时导致的连接中断问题
@@ -34,6 +85,117 @@ const keepAliveAgent = new https.Agent({
 
 logger.info('🌐 Gemini HTTPS Agent initialized with TCP Keep-Alive support')
 
+async function fetchAvailableModelsAntigravity(
+  accessToken,
+  proxyConfig = null,
+  refreshToken = null
+) {
+  try {
+    let effectiveToken = accessToken
+    if (refreshToken) {
+      try {
+        const client = await getOauthClient(
+          accessToken,
+          refreshToken,
+          proxyConfig,
+          OAUTH_PROVIDER_ANTIGRAVITY
+        )
+        if (client && client.getAccessToken) {
+          const latest = await client.getAccessToken()
+          if (latest?.token) {
+            effectiveToken = latest.token
+          }
+        }
+      } catch (error) {
+        logger.warn('Failed to refresh Antigravity access token for models list:', {
+          message: error.message
+        })
+      }
+    }
+
+    const data = await antigravityClient.fetchAvailableModels({
+      accessToken: effectiveToken,
+      proxyConfig
+    })
+    const modelsDict = data?.models
+    const created = Math.floor(Date.now() / 1000)
+
+    const models = []
+    const seen = new Set()
+    const {
+      getAntigravityModelAlias,
+      getAntigravityModelMetadata,
+      normalizeAntigravityModelInput
+    } = require('../utils/antigravityModel')
+
+    const pushModel = (modelId) => {
+      if (!modelId || seen.has(modelId)) {
+        return
+      }
+      seen.add(modelId)
+      const metadata = getAntigravityModelMetadata(modelId)
+      const entry = {
+        id: modelId,
+        object: 'model',
+        created,
+        owned_by: 'antigravity'
+      }
+      if (metadata?.name) {
+        entry.name = metadata.name
+      }
+      if (metadata?.maxCompletionTokens) {
+        entry.max_completion_tokens = metadata.maxCompletionTokens
+      }
+      if (metadata?.thinking) {
+        entry.thinking = metadata.thinking
+      }
+      models.push(entry)
+    }
+
+    if (modelsDict && typeof modelsDict === 'object') {
+      for (const modelId of Object.keys(modelsDict)) {
+        const normalized = normalizeAntigravityModelInput(modelId)
+        const alias = getAntigravityModelAlias(normalized)
+        if (!alias) {
+          continue
+        }
+        pushModel(alias)
+
+        if (alias.endsWith('-thinking')) {
+          pushModel(alias.replace(/-thinking$/, ''))
+        }
+
+        if (alias.startsWith('gemini-claude-')) {
+          pushModel(alias.replace(/^gemini-/, ''))
+        }
+      }
+    }
+
+    return models
+  } catch (error) {
+    logger.error('Failed to fetch Antigravity models:', error.response?.data || error.message)
+    return [
+      {
+        id: 'gemini-2.5-flash',
+        object: 'model',
+        created: Math.floor(Date.now() / 1000),
+        owned_by: 'antigravity'
+      }
+    ]
+  }
+}
+
+async function countTokensAntigravity(client, contents, model, proxyConfig = null) {
+  const { token } = await client.getAccessToken()
+  const response = await antigravityClient.countTokens({
+    accessToken: token,
+    proxyConfig,
+    contents,
+    model
+  })
+  return response
+}
+
 // 加密相关常量
 const ALGORITHM = 'aes-256-cbc'
 const ENCRYPTION_SALT = 'gemini-account-salt'
@@ -124,14 +286,15 @@ setInterval(
 )
 
 // 创建 OAuth2 客户端（支持代理配置）
-function createOAuth2Client(redirectUri = null, proxyConfig = null) {
+function createOAuth2Client(redirectUri = null, proxyConfig = null, oauthProvider = null) {
   // 如果没有提供 redirectUri，使用默认值
   const uri = redirectUri || 'http://localhost:45462'
+  const oauthConfig = getOauthProviderConfig(oauthProvider)
 
   // 准备客户端选项
   const clientOptions = {
-    clientId: OAUTH_CLIENT_ID,
-    clientSecret: OAUTH_CLIENT_SECRET,
+    clientId: oauthConfig.clientId,
+    clientSecret: oauthConfig.clientSecret,
     redirectUri: uri
   }
 
@@ -152,10 +315,17 @@ function createOAuth2Client(redirectUri = null, proxyConfig = null) {
 }
 
 // 生成授权 URL (支持 PKCE 和代理)
-async function generateAuthUrl(state = null, redirectUri = null, proxyConfig = null) {
+async function generateAuthUrl(
+  state = null,
+  redirectUri = null,
+  proxyConfig = null,
+  oauthProvider = null
+) {
   // 使用新的 redirect URI
   const finalRedirectUri = redirectUri || 'https://codeassist.google.com/authcode'
-  const oAuth2Client = createOAuth2Client(finalRedirectUri, proxyConfig)
+  const normalizedProvider = normalizeOauthProvider(oauthProvider)
+  const oauthConfig = getOauthProviderConfig(normalizedProvider)
+  const oAuth2Client = createOAuth2Client(finalRedirectUri, proxyConfig, normalizedProvider)
 
   if (proxyConfig) {
     logger.info(
@@ -172,7 +342,7 @@ async function generateAuthUrl(state = null, redirectUri = null, proxyConfig = n
   const authUrl = oAuth2Client.generateAuthUrl({
     redirect_uri: finalRedirectUri,
     access_type: 'offline',
-    scope: OAUTH_SCOPES,
+    scope: oauthConfig.scopes,
     code_challenge_method: 'S256',
     code_challenge: codeVerifier.codeChallenge,
     state: stateValue,
@@ -183,7 +353,8 @@ async function generateAuthUrl(state = null, redirectUri = null, proxyConfig = n
     authUrl,
     state: stateValue,
     codeVerifier: codeVerifier.codeVerifier,
-    redirectUri: finalRedirectUri
+    redirectUri: finalRedirectUri,
+    oauthProvider: normalizedProvider
   }
 }
 
@@ -244,11 +415,14 @@ async function exchangeCodeForTokens(
   code,
   redirectUri = null,
   codeVerifier = null,
-  proxyConfig = null
+  proxyConfig = null,
+  oauthProvider = null
 ) {
   try {
+    const normalizedProvider = normalizeOauthProvider(oauthProvider)
+    const oauthConfig = getOauthProviderConfig(normalizedProvider)
     // 创建带代理配置的 OAuth2Client
-    const oAuth2Client = createOAuth2Client(redirectUri, proxyConfig)
+    const oAuth2Client = createOAuth2Client(redirectUri, proxyConfig, normalizedProvider)
 
     if (proxyConfig) {
       logger.info(
@@ -274,7 +448,7 @@ async function exchangeCodeForTokens(
     return {
       access_token: tokens.access_token,
       refresh_token: tokens.refresh_token,
-      scope: tokens.scope || OAUTH_SCOPES.join(' '),
+      scope: tokens.scope || oauthConfig.scopes.join(' '),
       token_type: tokens.token_type || 'Bearer',
       expiry_date: tokens.expiry_date || Date.now() + tokens.expires_in * 1000
     }
@@ -285,9 +459,11 @@ async function exchangeCodeForTokens(
 }
 
 // 刷新访问令牌
-async function refreshAccessToken(refreshToken, proxyConfig = null) {
+async function refreshAccessToken(refreshToken, proxyConfig = null, oauthProvider = null) {
+  const normalizedProvider = normalizeOauthProvider(oauthProvider)
+  const oauthConfig = getOauthProviderConfig(normalizedProvider)
   // 创建带代理配置的 OAuth2Client
-  const oAuth2Client = createOAuth2Client(null, proxyConfig)
+  const oAuth2Client = createOAuth2Client(null, proxyConfig, normalizedProvider)
 
   try {
     // 设置 refresh_token
@@ -319,7 +495,7 @@ async function refreshAccessToken(refreshToken, proxyConfig = null) {
     return {
       access_token: credentials.access_token,
       refresh_token: credentials.refresh_token || refreshToken, // 保留原 refresh_token 如果没有返回新的
-      scope: credentials.scope || OAUTH_SCOPES.join(' '),
+      scope: credentials.scope || oauthConfig.scopes.join(' '),
       token_type: credentials.token_type || 'Bearer',
       expiry_date: credentials.expiry_date || Date.now() + 3600000 // 默认1小时过期
     }
@@ -339,6 +515,8 @@ async function refreshAccessToken(refreshToken, proxyConfig = null) {
 async function createAccount(accountData) {
   const id = uuidv4()
   const now = new Date().toISOString()
+  const oauthProvider = normalizeOauthProvider(accountData.oauthProvider)
+  const oauthConfig = getOauthProviderConfig(oauthProvider)
 
   // 处理凭证数据
   let geminiOauth = null
@@ -371,7 +549,7 @@ async function createAccount(accountData) {
       geminiOauth = JSON.stringify({
         access_token: accessToken,
         refresh_token: refreshToken,
-        scope: accountData.scope || OAUTH_SCOPES.join(' '),
+        scope: accountData.scope || oauthConfig.scopes.join(' '),
         token_type: accountData.tokenType || 'Bearer',
         expiry_date: accountData.expiryDate || Date.now() + 3600000 // 默认1小时
       })
@@ -399,7 +577,8 @@ async function createAccount(accountData) {
     refreshToken: refreshToken ? encrypt(refreshToken) : '',
     expiresAt, // OAuth Token 过期时间（技术字段，自动刷新）
     // 只有OAuth方式才有scopes，手动添加的没有
-    scopes: accountData.geminiOauth ? accountData.scopes || OAUTH_SCOPES.join(' ') : '',
+    scopes: accountData.geminiOauth ? accountData.scopes || oauthConfig.scopes.join(' ') : '',
+    oauthProvider,
 
     // ✅ 新增：账户订阅到期时间（业务字段，手动管理）
     subscriptionExpiresAt: accountData.subscriptionExpiresAt || null,
@@ -508,6 +687,10 @@ async function updateAccount(accountId, updates) {
     updates.schedulable = updates.schedulable.toString()
   }
 
+  if (updates.oauthProvider !== undefined) {
+    updates.oauthProvider = normalizeOauthProvider(updates.oauthProvider)
+  }
+
   // 加密敏感字段
   if (updates.geminiOauth) {
     updates.geminiOauth = encrypt(
@@ -885,12 +1068,13 @@ async function refreshAccountToken(accountId) {
       // 重新获取账户数据（可能已被其他进程刷新）
       const updatedAccount = await getAccount(accountId)
       if (updatedAccount && updatedAccount.accessToken) {
+        const oauthConfig = getOauthProviderConfig(updatedAccount.oauthProvider)
         const accessToken = decrypt(updatedAccount.accessToken)
         return {
           access_token: accessToken,
           refresh_token: updatedAccount.refreshToken ? decrypt(updatedAccount.refreshToken) : '',
           expiry_date: updatedAccount.expiresAt ? new Date(updatedAccount.expiresAt).getTime() : 0,
-          scope: updatedAccount.scope || OAUTH_SCOPES.join(' '),
+          scope: updatedAccount.scopes || oauthConfig.scopes.join(' '),
           token_type: 'Bearer'
         }
       }
@@ -904,7 +1088,11 @@ async function refreshAccountToken(accountId) {
 
     // account.refreshToken 已经是解密后的值（从 getAccount 返回）
     // 传入账户的代理配置
-    const newTokens = await refreshAccessToken(account.refreshToken, account.proxy)
+    const newTokens = await refreshAccessToken(
+      account.refreshToken,
+      account.proxy,
+      account.oauthProvider
+    )
 
     // 更新账户信息
     const updates = {
@@ -1036,14 +1224,15 @@ async function getAccountRateLimitInfo(accountId) {
 }
 
 // 获取配置的OAuth客户端 - 参考GeminiCliSimulator的getOauthClient方法（支持代理）
-async function getOauthClient(accessToken, refreshToken, proxyConfig = null) {
-  const client = createOAuth2Client(null, proxyConfig)
+async function getOauthClient(accessToken, refreshToken, proxyConfig = null, oauthProvider = null) {
+  const normalizedProvider = normalizeOauthProvider(oauthProvider)
+  const oauthConfig = getOauthProviderConfig(normalizedProvider)
+  const client = createOAuth2Client(null, proxyConfig, normalizedProvider)
 
   const creds = {
     access_token: accessToken,
     refresh_token: refreshToken,
-    scope:
-      'https://www.googleapis.com/auth/cloud-platform https://www.googleapis.com/auth/userinfo.profile openid https://www.googleapis.com/auth/userinfo.email',
+    scope: oauthConfig.scopes.join(' '),
     token_type: 'Bearer',
     expiry_date: 1754269905646
   }
@@ -1509,6 +1698,43 @@ async function generateContent(
   return response.data
 }
 
+// 调用 Antigravity 上游生成内容（非流式）
+async function generateContentAntigravity(
+  client,
+  requestData,
+  userPromptId,
+  projectId = null,
+  sessionId = null,
+  proxyConfig = null
+) {
+  const { token } = await client.getAccessToken()
+  const { model } = antigravityClient.buildAntigravityEnvelope({
+    requestData,
+    projectId,
+    sessionId,
+    userPromptId
+  })
+
+  logger.info('🪐 Antigravity generateContent API调用开始', {
+    model,
+    userPromptId,
+    projectId,
+    sessionId
+  })
+
+  const { response } = await antigravityClient.request({
+    accessToken: token,
+    proxyConfig,
+    requestData,
+    projectId,
+    sessionId,
+    userPromptId,
+    stream: false
+  })
+  logger.info('✅ Antigravity generateContent API调用成功')
+  return response.data
+}
+
 // 调用 Code Assist API 生成内容（流式）
 async function generateContentStream(
   client,
@@ -1593,6 +1819,46 @@ async function generateContentStream(
   return response.data // 返回流对象
 }
 
+// 调用 Antigravity 上游生成内容（流式）
+async function generateContentStreamAntigravity(
+  client,
+  requestData,
+  userPromptId,
+  projectId = null,
+  sessionId = null,
+  signal = null,
+  proxyConfig = null
+) {
+  const { token } = await client.getAccessToken()
+  const { model } = antigravityClient.buildAntigravityEnvelope({
+    requestData,
+    projectId,
+    sessionId,
+    userPromptId
+  })
+
+  logger.info('🌊 Antigravity streamGenerateContent API调用开始', {
+    model,
+    userPromptId,
+    projectId,
+    sessionId
+  })
+
+  const { response } = await antigravityClient.request({
+    accessToken: token,
+    proxyConfig,
+    requestData,
+    projectId,
+    sessionId,
+    userPromptId,
+    stream: true,
+    signal,
+    params: { alt: 'sse' }
+  })
+  logger.info('✅ Antigravity streamGenerateContent API调用成功，开始流式传输')
+  return response.data
+}
+
 // 更新账户的临时项目 ID
 async function updateTempProjectId(accountId, tempProjectId) {
   if (!tempProjectId) {
@@ -1687,10 +1953,12 @@ module.exports = {
   generateEncryptionKey,
   decryptCache, // 暴露缓存对象以便测试和监控
   countTokens,
+  countTokensAntigravity,
   generateContent,
   generateContentStream,
+  generateContentAntigravity,
+  generateContentStreamAntigravity,
+  fetchAvailableModelsAntigravity,
   updateTempProjectId,
-  resetAccountStatus,
-  OAUTH_CLIENT_ID,
-  OAUTH_SCOPES
+  resetAccountStatus
 }

src/services/openaiResponsesRelayService.js

@@ -1,6 +1,7 @@
 const axios = require('axios')
 const ProxyHelper = require('../utils/proxyHelper')
 const logger = require('../utils/logger')
+const { filterForOpenAI } = require('../utils/headerFilter')
 const openaiResponsesAccountService = require('./openaiResponsesAccountService')
 const apiKeyService = require('./apiKeyService')
 const unifiedOpenAIScheduler = require('./unifiedOpenAIScheduler')
@@ -73,9 +74,9 @@ class OpenAIResponsesRelayService {
       const targetUrl = `${fullAccount.baseApi}${req.path}`
       logger.info(`🎯 Forwarding to: ${targetUrl}`)
 
-      // 构建请求头
+      // 构建请求头 - 使用统一的 headerFilter 移除 CDN headers
       const headers = {
-        ...this._filterRequestHeaders(req.headers),
+        ...filterForOpenAI(req.headers),
         Authorization: `Bearer ${fullAccount.apiKey}`,
         'Content-Type': 'application/json'
       }
@@ -425,9 +426,9 @@ class OpenAIResponsesRelayService {
       const lines = data.split('\n')
 
       for (const line of lines) {
-        if (line.startsWith('data: ')) {
+        if (line.startsWith('data:')) {
           try {
-            const jsonStr = line.slice(6)
+            const jsonStr = line.slice(5).trim()
             if (jsonStr === '[DONE]') {
               continue
             }
@@ -810,29 +811,10 @@ class OpenAIResponsesRelayService {
     return { resetsInSeconds, errorData }
   }
 
-  // 过滤请求头
+  // 过滤请求头 - 已迁移到 headerFilter 工具类
+  // 此方法保留用于向后兼容，实际使用 filterForOpenAI()
   _filterRequestHeaders(headers) {
-    const filtered = {}
-    const skipHeaders = [
-      'host',
-      'content-length',
-      'authorization',
-      'x-api-key',
-      'x-cr-api-key',
-      'connection',
-      'upgrade',
-      'sec-websocket-key',
-      'sec-websocket-version',
-      'sec-websocket-extensions'
-    ]
-
-    for (const [key, value] of Object.entries(headers)) {
-      if (!skipHeaders.includes(key.toLowerCase())) {
-        filtered[key] = value
-      }
-    }
-
-    return filtered
+    return filterForOpenAI(headers)
   }
 
   // 估算费用（简化版本，实际应该根据不同的定价模型）

src/services/rateLimitCleanupService.js

@@ -72,7 +72,8 @@ class RateLimitCleanupService {
       const results = {
         openai: { checked: 0, cleared: 0, errors: [] },
         claude: { checked: 0, cleared: 0, errors: [] },
-        claudeConsole: { checked: 0, cleared: 0, errors: [] }
+        claudeConsole: { checked: 0, cleared: 0, errors: [] },
+        tokenRefresh: { checked: 0, refreshed: 0, errors: [] }
       }
 
       // 清理 OpenAI 账号
@@ -84,21 +85,29 @@ class RateLimitCleanupService {
       // 清理 Claude Console 账号
       await this.cleanupClaudeConsoleAccounts(results.claudeConsole)
 
+      // 主动刷新等待重置的 Claude 账户 Token（防止 5小时/7天 等待期间 Token 过期）
+      await this.proactiveRefreshClaudeTokens(results.tokenRefresh)
+
       const totalChecked =
         results.openai.checked + results.claude.checked + results.claudeConsole.checked
       const totalCleared =
         results.openai.cleared + results.claude.cleared + results.claudeConsole.cleared
       const duration = Date.now() - startTime
 
-      if (totalCleared > 0) {
+      if (totalCleared > 0 || results.tokenRefresh.refreshed > 0) {
         logger.info(
-          `✅ Rate limit cleanup completed: ${totalCleared} accounts cleared out of ${totalChecked} checked (${duration}ms)`
+          `✅ Rate limit cleanup completed: ${totalCleared}/${totalChecked} accounts cleared, ${results.tokenRefresh.refreshed} tokens refreshed (${duration}ms)`
         )
         logger.info(`   OpenAI: ${results.openai.cleared}/${results.openai.checked}`)
         logger.info(`   Claude: ${results.claude.cleared}/${results.claude.checked}`)
         logger.info(
           `   Claude Console: ${results.claudeConsole.cleared}/${results.claudeConsole.checked}`
         )
+        if (results.tokenRefresh.checked > 0 || results.tokenRefresh.refreshed > 0) {
+          logger.info(
+            `   Token Refresh: ${results.tokenRefresh.refreshed}/${results.tokenRefresh.checked} refreshed`
+          )
+        }
 
         // 发送 webhook 恢复通知
         if (this.clearedAccounts.length > 0) {
@@ -114,7 +123,8 @@ class RateLimitCleanupService {
       const allErrors = [
         ...results.openai.errors,
         ...results.claude.errors,
-        ...results.claudeConsole.errors
+        ...results.claudeConsole.errors,
+        ...results.tokenRefresh.errors
       ]
       if (allErrors.length > 0) {
         logger.warn(`⚠️ Encountered ${allErrors.length} errors during cleanup:`, allErrors)
@@ -348,6 +358,75 @@ class RateLimitCleanupService {
     }
   }
 
+  /**
+   * 主动刷新 Claude 账户 Token（防止等待重置期间 Token 过期）
+   * 仅对因限流/配额限制而等待重置的账户执行刷新：
+   * - 429 限流账户（rateLimitAutoStopped=true）
+   * - 5小时限制自动停止账户（fiveHourAutoStopped=true）
+   * 不处理错误状态账户（error/temp_error）
+   */
+  async proactiveRefreshClaudeTokens(result) {
+    try {
+      const redis = require('../models/redis')
+      const accounts = await redis.getAllClaudeAccounts()
+      const now = Date.now()
+      const refreshAheadMs = 30 * 60 * 1000 // 提前30分钟刷新
+      const recentRefreshMs = 5 * 60 * 1000 // 5分钟内刷新过则跳过
+
+      for (const account of accounts) {
+        // 1. 必须激活
+        if (account.isActive !== 'true') {
+          continue
+        }
+
+        // 2. 必须有 refreshToken
+        if (!account.refreshToken) {
+          continue
+        }
+
+        // 3. 【优化】仅处理因限流/配额限制而等待重置的账户
+        // 正常调度的账户会在请求时自动刷新，无需主动刷新
+        // 错误状态账户的 Token 可能已失效，刷新也会失败
+        const isWaitingForReset =
+          account.rateLimitAutoStopped === 'true' || // 429 限流
+          account.fiveHourAutoStopped === 'true' // 5小时限制自动停止
+        if (!isWaitingForReset) {
+          continue
+        }
+
+        // 4. 【优化】如果最近 5 分钟内已刷新，跳过（避免重复刷新）
+        const lastRefreshAt = account.lastRefreshAt ? new Date(account.lastRefreshAt).getTime() : 0
+        if (now - lastRefreshAt < recentRefreshMs) {
+          continue
+        }
+
+        // 5. 检查 Token 是否即将过期（30分钟内）
+        const expiresAt = parseInt(account.expiresAt)
+        if (expiresAt && now < expiresAt - refreshAheadMs) {
+          continue
+        }
+
+        // 符合条件，执行刷新
+        result.checked++
+        try {
+          await claudeAccountService.refreshAccountToken(account.id)
+          result.refreshed++
+          logger.info(`🔄 Proactively refreshed token: ${account.name} (${account.id})`)
+        } catch (error) {
+          result.errors.push({
+            accountId: account.id,
+            accountName: account.name,
+            error: error.message
+          })
+          logger.warn(`⚠️ Proactive refresh failed for ${account.name}: ${error.message}`)
+        }
+      }
+    } catch (error) {
+      logger.error('Failed to proactively refresh Claude tokens:', error)
+      result.errors.push({ error: error.message })
+    }
+  }
+
   /**
    * 手动触发一次清理（供 API 或 CLI 调用）
    */

src/services/requestIdentityService.js

@@ -22,6 +22,18 @@ const STAINLESS_HEADER_KEYS = [
   'x-stainless-runtime',
   'x-stainless-runtime-version'
 ]
+
+// 小写 key 到正确大小写格式的映射（用于返回给上游时）
+const STAINLESS_HEADER_CASE_MAP = {
+  'x-stainless-retry-count': 'X-Stainless-Retry-Count',
+  'x-stainless-timeout': 'X-Stainless-Timeout',
+  'x-stainless-lang': 'X-Stainless-Lang',
+  'x-stainless-package-version': 'X-Stainless-Package-Version',
+  'x-stainless-os': 'X-Stainless-OS',
+  'x-stainless-arch': 'X-Stainless-Arch',
+  'x-stainless-runtime': 'X-Stainless-Runtime',
+  'x-stainless-runtime-version': 'X-Stainless-Runtime-Version'
+}
 const MIN_FINGERPRINT_FIELDS = 4
 const REDIS_KEY_PREFIX = 'fmt_claude_req:stainless_headers:'
 
@@ -135,7 +147,9 @@ function applyFingerprintToHeaders(headers, fingerprint) {
       return
     }
     removeHeaderCaseInsensitive(nextHeaders, key)
-    nextHeaders[key] = fingerprint[key]
+    // 使用正确的大小写格式返回给上游
+    const properCaseKey = STAINLESS_HEADER_CASE_MAP[key] || key
+    nextHeaders[properCaseKey] = fingerprint[key]
   })
 
   return nextHeaders

src/services/unifiedClaudeScheduler.js

@@ -5,7 +5,33 @@ const ccrAccountService = require('./ccrAccountService')
 const accountGroupService = require('./accountGroupService')
 const redis = require('../models/redis')
 const logger = require('../utils/logger')
-const { parseVendorPrefixedModel } = require('../utils/modelHelper')
+const { parseVendorPrefixedModel, isOpus45OrNewer } = require('../utils/modelHelper')
+
+/**
+ * Check if account is Pro (not Max)
+ *
+ * ACCOUNT TYPE LOGIC (as of 2025-12-05):
+ * Pro accounts can be identified by either:
+ *   1. API real-time data: hasClaudePro=true && hasClaudeMax=false
+ *   2. Local config data: accountType='claude_pro'
+ *
+ * Account type restrictions for Opus models:
+ *   - Free account: No Opus access at all
+ *   - Pro account: Only Opus 4.5+ (new versions)
+ *   - Max account: All Opus versions (legacy 3.x, 4.0, 4.1 and new 4.5+)
+ *
+ * Compatible with both API real-time data (hasClaudePro) and local config (accountType)
+ * @param {Object} info - Subscription info object
+ * @returns {boolean} - true if Pro account (not Free, not Max)
+ */
+function isProAccount(info) {
+  // API real-time status takes priority
+  if (info.hasClaudePro === true && info.hasClaudeMax !== true) {
+    return true
+  }
+  // Local configured account type
+  return info.accountType === 'claude_pro'
+}
 
 class UnifiedClaudeScheduler {
   constructor() {
@@ -46,8 +72,14 @@ class UnifiedClaudeScheduler {
         return false
       }
 
-      // 2. Opus 模型的订阅级别检查
+      // 2. Opus model subscription level check
+      // VERSION RESTRICTION LOGIC:
+      // - Free: No Opus models
+      // - Pro: Only Opus 4.5+ (isOpus45OrNewer = true)
+      // - Max: All Opus versions
       if (requestedModel.toLowerCase().includes('opus')) {
+        const isNewOpus = isOpus45OrNewer(requestedModel)
+
         if (account.subscriptionInfo) {
           try {
             const info =
@@ -55,27 +87,36 @@ class UnifiedClaudeScheduler {
                 ? JSON.parse(account.subscriptionInfo)
                 : account.subscriptionInfo
 
-            // Pro 和 Free 账号不支持 Opus
-            if (info.hasClaudePro === true && info.hasClaudeMax !== true) {
+            // Free account: does not support any Opus model
+            if (info.accountType === 'free') {
               logger.info(
-                `🚫 Claude account ${account.name} (Pro) does not support Opus model${context ? ` ${context}` : ''}`
+                `🚫 Claude account ${account.name} (Free) does not support Opus model${context ? ` ${context}` : ''}`
               )
               return false
             }
-            if (info.accountType === 'claude_pro' || info.accountType === 'claude_free') {
-              logger.info(
-                `🚫 Claude account ${account.name} (${info.accountType}) does not support Opus model${context ? ` ${context}` : ''}`
-              )
-              return false
+
+            // Pro account: only supports Opus 4.5+
+            // Reject legacy Opus (3.x, 4.0-4.4) but allow new Opus (4.5+)
+            if (isProAccount(info)) {
+              if (!isNewOpus) {
+                logger.info(
+                  `🚫 Claude account ${account.name} (Pro) does not support legacy Opus model${context ? ` ${context}` : ''}`
+                )
+                return false
+              }
+              // Opus 4.5+ supported
+              return true
             }
+
+            // Max account: supports all Opus versions (no restriction)
           } catch (e) {
-            // 解析失败，假设为旧数据，默认支持（兼容旧数据为 Max）
+            // Parse failed, assume legacy data (Max), default support
             logger.debug(
               `Account ${account.name} has invalid subscriptionInfo${context ? ` ${context}` : ''}, assuming Max`
             )
           }
         }
-        // 没有订阅信息的账号，默认当作支持（兼容旧数据）
+        // Account without subscription info, default to supported (legacy data compatibility)
       }
     }
 
@@ -139,8 +180,56 @@ class UnifiedClaudeScheduler {
   }
 
   // 🎯 统一调度Claude账号（官方和Console）
-  async selectAccountForApiKey(apiKeyData, sessionHash = null, requestedModel = null) {
+  async selectAccountForApiKey(
+    apiKeyData,
+    sessionHash = null,
+    requestedModel = null,
+    forcedAccount = null
+  ) {
     try {
+      // 🔒 如果有强制绑定的账户（全局会话绑定），仅 claude-official 类型受影响
+      if (forcedAccount && forcedAccount.accountId && forcedAccount.accountType) {
+        // ⚠️ 只有 claude-official 类型账户受全局会话绑定限制
+        // 其他类型（bedrock, ccr, claude-console等）忽略绑定，走正常调度
+        if (forcedAccount.accountType !== 'claude-official') {
+          logger.info(
+            `🔗 Session binding ignored for non-official account type: ${forcedAccount.accountType}, proceeding with normal scheduling`
+          )
+          // 不使用 forcedAccount，继续走下面的正常调度逻辑
+        } else {
+          // claude-official 类型需要检查可用性并强制使用
+          logger.info(
+            `🔗 Forced session binding detected: ${forcedAccount.accountId} (${forcedAccount.accountType})`
+          )
+
+          const isAvailable = await this._isAccountAvailableForSessionBinding(
+            forcedAccount.accountId,
+            forcedAccount.accountType,
+            requestedModel
+          )
+
+          if (isAvailable) {
+            logger.info(
+              `✅ Using forced session binding account: ${forcedAccount.accountId} (${forcedAccount.accountType})`
+            )
+            return {
+              accountId: forcedAccount.accountId,
+              accountType: forcedAccount.accountType
+            }
+          } else {
+            // 绑定账户不可用，抛出特定错误（不 fallback）
+            logger.warn(
+              `❌ Forced session binding account unavailable: ${forcedAccount.accountId} (${forcedAccount.accountType})`
+            )
+            const error = new Error('Session binding account unavailable')
+            error.code = 'SESSION_BINDING_ACCOUNT_UNAVAILABLE'
+            error.accountId = forcedAccount.accountId
+            error.accountType = forcedAccount.accountType
+            throw error
+          }
+        }
+      }
+
       // 解析供应商前缀
       const { vendor, baseModel } = parseVendorPrefixedModel(requestedModel)
       const effectiveModel = vendor === 'ccr' ? baseModel : requestedModel
@@ -177,30 +266,41 @@ class UnifiedClaudeScheduler {
         // 普通专属账户
         const boundAccount = await redis.getClaudeAccount(apiKeyData.claudeAccountId)
         if (boundAccount && boundAccount.isActive === 'true' && boundAccount.status !== 'error') {
-          const isRateLimited = await claudeAccountService.isAccountRateLimited(boundAccount.id)
-          if (isRateLimited) {
-            const rateInfo = await claudeAccountService.getAccountRateLimitInfo(boundAccount.id)
-            const error = new Error('Dedicated Claude account is rate limited')
-            error.code = 'CLAUDE_DEDICATED_RATE_LIMITED'
-            error.accountId = boundAccount.id
-            error.rateLimitEndAt = rateInfo?.rateLimitEndAt || boundAccount.rateLimitEndAt || null
-            throw error
-          }
-
-          if (!this._isSchedulable(boundAccount.schedulable)) {
+          // 检查是否临时不可用
+          const isTempUnavailable = await this.isAccountTemporarilyUnavailable(
+            boundAccount.id,
+            'claude-official'
+          )
+          if (isTempUnavailable) {
             logger.warn(
-              `⚠️ Bound Claude OAuth account ${apiKeyData.claudeAccountId} is not schedulable (schedulable: ${boundAccount?.schedulable}), falling back to pool`
+              `⏱️ Bound Claude OAuth account ${boundAccount.id} is temporarily unavailable, falling back to pool`
             )
           } else {
-            if (isOpusRequest) {
-              await claudeAccountService.clearExpiredOpusRateLimit(boundAccount.id)
+            const isRateLimited = await claudeAccountService.isAccountRateLimited(boundAccount.id)
+            if (isRateLimited) {
+              const rateInfo = await claudeAccountService.getAccountRateLimitInfo(boundAccount.id)
+              const error = new Error('Dedicated Claude account is rate limited')
+              error.code = 'CLAUDE_DEDICATED_RATE_LIMITED'
+              error.accountId = boundAccount.id
+              error.rateLimitEndAt = rateInfo?.rateLimitEndAt || boundAccount.rateLimitEndAt || null
+              throw error
             }
-            logger.info(
-              `🎯 Using bound dedicated Claude OAuth account: ${boundAccount.name} (${apiKeyData.claudeAccountId}) for API key ${apiKeyData.name}`
-            )
-            return {
-              accountId: apiKeyData.claudeAccountId,
-              accountType: 'claude-official'
+
+            if (!this._isSchedulable(boundAccount.schedulable)) {
+              logger.warn(
+                `⚠️ Bound Claude OAuth account ${apiKeyData.claudeAccountId} is not schedulable (schedulable: ${boundAccount?.schedulable}), falling back to pool`
+              )
+            } else {
+              if (isOpusRequest) {
+                await claudeAccountService.clearExpiredOpusRateLimit(boundAccount.id)
+              }
+              logger.info(
+                `🎯 Using bound dedicated Claude OAuth account: ${boundAccount.name} (${apiKeyData.claudeAccountId}) for API key ${apiKeyData.name}`
+              )
+              return {
+                accountId: apiKeyData.claudeAccountId,
+                accountType: 'claude-official'
+              }
             }
           }
         } else {
@@ -221,12 +321,23 @@ class UnifiedClaudeScheduler {
           boundConsoleAccount.status === 'active' &&
           this._isSchedulable(boundConsoleAccount.schedulable)
         ) {
-          logger.info(
-            `🎯 Using bound dedicated Claude Console account: ${boundConsoleAccount.name} (${apiKeyData.claudeConsoleAccountId}) for API key ${apiKeyData.name}`
+          // 检查是否临时不可用
+          const isTempUnavailable = await this.isAccountTemporarilyUnavailable(
+            boundConsoleAccount.id,
+            'claude-console'
           )
-          return {
-            accountId: apiKeyData.claudeConsoleAccountId,
-            accountType: 'claude-console'
+          if (isTempUnavailable) {
+            logger.warn(
+              `⏱️ Bound Claude Console account ${boundConsoleAccount.id} is temporarily unavailable, falling back to pool`
+            )
+          } else {
+            logger.info(
+              `🎯 Using bound dedicated Claude Console account: ${boundConsoleAccount.name} (${apiKeyData.claudeConsoleAccountId}) for API key ${apiKeyData.name}`
+            )
+            return {
+              accountId: apiKeyData.claudeConsoleAccountId,
+              accountType: 'claude-console'
+            }
           }
         } else {
           logger.warn(
@@ -245,12 +356,23 @@ class UnifiedClaudeScheduler {
           boundBedrockAccountResult.data.isActive === true &&
           this._isSchedulable(boundBedrockAccountResult.data.schedulable)
         ) {
-          logger.info(
-            `🎯 Using bound dedicated Bedrock account: ${boundBedrockAccountResult.data.name} (${apiKeyData.bedrockAccountId}) for API key ${apiKeyData.name}`
+          // 检查是否临时不可用
+          const isTempUnavailable = await this.isAccountTemporarilyUnavailable(
+            apiKeyData.bedrockAccountId,
+            'bedrock'
           )
-          return {
-            accountId: apiKeyData.bedrockAccountId,
-            accountType: 'bedrock'
+          if (isTempUnavailable) {
+            logger.warn(
+              `⏱️ Bound Bedrock account ${apiKeyData.bedrockAccountId} is temporarily unavailable, falling back to pool`
+            )
+          } else {
+            logger.info(
+              `🎯 Using bound dedicated Bedrock account: ${boundBedrockAccountResult.data.name} (${apiKeyData.bedrockAccountId}) for API key ${apiKeyData.name}`
+            )
+            return {
+              accountId: apiKeyData.bedrockAccountId,
+              accountType: 'bedrock'
+            }
           }
         } else {
           logger.warn(
@@ -496,6 +618,18 @@ class UnifiedClaudeScheduler {
           continue
         }
 
+        // 检查是否临时不可用
+        const isTempUnavailable = await this.isAccountTemporarilyUnavailable(
+          account.id,
+          'claude-official'
+        )
+        if (isTempUnavailable) {
+          logger.debug(
+            `⏭️ Skipping Claude Official account ${account.name} - temporarily unavailable`
+          )
+          continue
+        }
+
         // 检查是否被限流
         const isRateLimited = await claudeAccountService.isAccountRateLimited(account.id)
         if (isRateLimited) {
@@ -584,6 +718,18 @@ class UnifiedClaudeScheduler {
           // 继续处理该账号
         }
 
+        // 检查是否临时不可用
+        const isTempUnavailable = await this.isAccountTemporarilyUnavailable(
+          currentAccount.id,
+          'claude-console'
+        )
+        if (isTempUnavailable) {
+          logger.debug(
+            `⏭️ Skipping Claude Console account ${currentAccount.name} - temporarily unavailable`
+          )
+          continue
+        }
+
         // 检查是否被限流
         const isRateLimited = await claudeConsoleAccountService.isAccountRateLimited(
           currentAccount.id
@@ -682,7 +828,15 @@ class UnifiedClaudeScheduler {
           account.accountType === 'shared' &&
           this._isSchedulable(account.schedulable)
         ) {
-          // 检查是否可调度
+          // 检查是否临时不可用
+          const isTempUnavailable = await this.isAccountTemporarilyUnavailable(
+            account.id,
+            'bedrock'
+          )
+          if (isTempUnavailable) {
+            logger.debug(`⏭️ Skipping Bedrock account ${account.name} - temporarily unavailable`)
+            continue
+          }
 
           availableAccounts.push({
             ...account,
@@ -731,6 +885,13 @@ class UnifiedClaudeScheduler {
             continue
           }
 
+          // 检查是否临时不可用
+          const isTempUnavailable = await this.isAccountTemporarilyUnavailable(account.id, 'ccr')
+          if (isTempUnavailable) {
+            logger.debug(`⏭️ Skipping CCR account ${account.name} - temporarily unavailable`)
+            continue
+          }
+
           // 检查是否被限流
           const isRateLimited = await ccrAccountService.isAccountRateLimited(account.id)
           const isQuotaExceeded = await ccrAccountService.isAccountQuotaExceeded(account.id)
@@ -1099,6 +1260,42 @@ class UnifiedClaudeScheduler {
     }
   }
 
+  // ⏱️ 标记账户为临时不可用状态（用于5xx等临时故障，默认5分钟后自动恢复）
+  async markAccountTemporarilyUnavailable(
+    accountId,
+    accountType,
+    sessionHash = null,
+    ttlSeconds = 300
+  ) {
+    try {
+      const client = redis.getClientSafe()
+      const key = `temp_unavailable:${accountType}:${accountId}`
+      await client.setex(key, ttlSeconds, '1')
+      if (sessionHash) {
+        await this._deleteSessionMapping(sessionHash)
+      }
+      logger.warn(
+        `⏱️ Account ${accountId} (${accountType}) marked temporarily unavailable for ${ttlSeconds}s`
+      )
+      return { success: true }
+    } catch (error) {
+      logger.error(`❌ Failed to mark account temporarily unavailable: ${accountId}`, error)
+      return { success: false }
+    }
+  }
+
+  // 🔍 检查账户是否临时不可用
+  async isAccountTemporarilyUnavailable(accountId, accountType) {
+    try {
+      const client = redis.getClientSafe()
+      const key = `temp_unavailable:${accountType}:${accountId}`
+      return (await client.exists(key)) === 1
+    } catch (error) {
+      logger.error(`❌ Failed to check temp unavailable status: ${accountId}`, error)
+      return false
+    }
+  }
+
   // 🚫 标记账户为限流状态
   async markAccountRateLimited(
     accountId,
@@ -1562,6 +1759,67 @@ class UnifiedClaudeScheduler {
       return []
     }
   }
+
+  /**
+   * 🔒 检查 claude-official 账户是否可用于会话绑定
+   * 注意：此方法仅用于 claude-official 类型账户，其他类型不受会话绑定限制
+   * @param {string} accountId - 账户ID
+   * @param {string} accountType - 账户类型（应为 'claude-official'）
+   * @param {string} _requestedModel - 请求的模型（保留参数，当前未使用）
+   * @returns {Promise<boolean>}
+   */
+  async _isAccountAvailableForSessionBinding(accountId, accountType, _requestedModel = null) {
+    try {
+      // 此方法仅处理 claude-official 类型
+      if (accountType !== 'claude-official') {
+        logger.warn(
+          `Session binding: _isAccountAvailableForSessionBinding called for non-official type: ${accountType}`
+        )
+        return true // 非 claude-official 类型不受限制
+      }
+
+      const account = await redis.getClaudeAccount(accountId)
+      if (!account) {
+        logger.warn(`Session binding: Claude OAuth account ${accountId} not found`)
+        return false
+      }
+
+      const isActive = account.isActive === 'true' || account.isActive === true
+      const { status } = account
+
+      if (!isActive) {
+        logger.warn(`Session binding: Claude OAuth account ${accountId} is not active`)
+        return false
+      }
+
+      if (status === 'error' || status === 'temp_error') {
+        logger.warn(
+          `Session binding: Claude OAuth account ${accountId} has error status: ${status}`
+        )
+        return false
+      }
+
+      // 检查是否被限流
+      if (await claudeAccountService.isAccountRateLimited(accountId)) {
+        logger.warn(`Session binding: Claude OAuth account ${accountId} is rate limited`)
+        return false
+      }
+
+      // 检查临时不可用
+      if (await this.isAccountTemporarilyUnavailable(accountId, accountType)) {
+        logger.warn(`Session binding: Claude OAuth account ${accountId} is temporarily unavailable`)
+        return false
+      }
+
+      return true
+    } catch (error) {
+      logger.error(
+        `❌ Error checking account availability for session binding: ${accountId} (${accountType})`,
+        error
+      )
+      return false
+    }
+  }
 }
 
 module.exports = new UnifiedClaudeScheduler()

src/services/unifiedGeminiScheduler.js

@@ -4,11 +4,35 @@ const accountGroupService = require('./accountGroupService')
 const redis = require('../models/redis')
 const logger = require('../utils/logger')
 
+const OAUTH_PROVIDER_GEMINI_CLI = 'gemini-cli'
+const OAUTH_PROVIDER_ANTIGRAVITY = 'antigravity'
+const KNOWN_OAUTH_PROVIDERS = [OAUTH_PROVIDER_GEMINI_CLI, OAUTH_PROVIDER_ANTIGRAVITY]
+
+function normalizeOauthProvider(oauthProvider) {
+  if (!oauthProvider) {
+    return OAUTH_PROVIDER_GEMINI_CLI
+  }
+  return oauthProvider === OAUTH_PROVIDER_ANTIGRAVITY
+    ? OAUTH_PROVIDER_ANTIGRAVITY
+    : OAUTH_PROVIDER_GEMINI_CLI
+}
+
 class UnifiedGeminiScheduler {
   constructor() {
     this.SESSION_MAPPING_PREFIX = 'unified_gemini_session_mapping:'
   }
 
+  _getSessionMappingKey(sessionHash, oauthProvider = null) {
+    if (!sessionHash) {
+      return null
+    }
+    if (!oauthProvider) {
+      return `${this.SESSION_MAPPING_PREFIX}${sessionHash}`
+    }
+    const normalized = normalizeOauthProvider(oauthProvider)
+    return `${this.SESSION_MAPPING_PREFIX}${normalized}:${sessionHash}`
+  }
+
   // 🔧 辅助方法：检查账户是否可调度（兼容字符串和布尔值）
   _isSchedulable(schedulable) {
     // 如果是 undefined 或 null，默认为可调度
@@ -32,7 +56,8 @@ class UnifiedGeminiScheduler {
     requestedModel = null,
     options = {}
   ) {
-    const { allowApiAccounts = false } = options
+    const { allowApiAccounts = false, oauthProvider = null } = options
+    const normalizedOauthProvider = oauthProvider ? normalizeOauthProvider(oauthProvider) : null
 
     try {
       // 如果API Key绑定了专属账户或分组，优先使用
@@ -83,14 +108,23 @@ class UnifiedGeminiScheduler {
             this._isActive(boundAccount.isActive) &&
             boundAccount.status !== 'error'
           ) {
-            logger.info(
-              `🎯 Using bound dedicated Gemini account: ${boundAccount.name} (${apiKeyData.geminiAccountId}) for API key ${apiKeyData.name}`
-            )
-            // 更新账户的最后使用时间
-            await geminiAccountService.markAccountUsed(apiKeyData.geminiAccountId)
-            return {
-              accountId: apiKeyData.geminiAccountId,
-              accountType: 'gemini'
+            if (
+              normalizedOauthProvider &&
+              normalizeOauthProvider(boundAccount.oauthProvider) !== normalizedOauthProvider
+            ) {
+              logger.warn(
+                `⚠️ Bound Gemini OAuth account ${boundAccount.name} oauthProvider=${normalizeOauthProvider(boundAccount.oauthProvider)} does not match requested oauthProvider=${normalizedOauthProvider}, falling back to pool`
+              )
+            } else {
+              logger.info(
+                `🎯 Using bound dedicated Gemini account: ${boundAccount.name} (${apiKeyData.geminiAccountId}) for API key ${apiKeyData.name}`
+              )
+              // 更新账户的最后使用时间
+              await geminiAccountService.markAccountUsed(apiKeyData.geminiAccountId)
+              return {
+                accountId: apiKeyData.geminiAccountId,
+                accountType: 'gemini'
+              }
             }
           } else {
             logger.warn(
@@ -102,7 +136,7 @@ class UnifiedGeminiScheduler {
 
       // 如果有会话哈希，检查是否有已映射的账户
       if (sessionHash) {
-        const mappedAccount = await this._getSessionMapping(sessionHash)
+        const mappedAccount = await this._getSessionMapping(sessionHash, normalizedOauthProvider)
         if (mappedAccount) {
           // 验证映射的账户是否仍然可用
           const isAvailable = await this._isAccountAvailable(
@@ -111,7 +145,7 @@ class UnifiedGeminiScheduler {
           )
           if (isAvailable) {
             // 🚀 智能会话续期（续期 unified 映射键，按配置）
-            await this._extendSessionMappingTTL(sessionHash)
+            await this._extendSessionMappingTTL(sessionHash, normalizedOauthProvider)
             logger.info(
               `🎯 Using sticky session account: ${mappedAccount.accountId} (${mappedAccount.accountType}) for session ${sessionHash}`
             )
@@ -132,11 +166,10 @@ class UnifiedGeminiScheduler {
       }
 
       // 获取所有可用账户
-      const availableAccounts = await this._getAllAvailableAccounts(
-        apiKeyData,
-        requestedModel,
-        allowApiAccounts
-      )
+      const availableAccounts = await this._getAllAvailableAccounts(apiKeyData, requestedModel, {
+        allowApiAccounts,
+        oauthProvider: normalizedOauthProvider
+      })
 
       if (availableAccounts.length === 0) {
         // 提供更详细的错误信息
@@ -160,7 +193,8 @@ class UnifiedGeminiScheduler {
         await this._setSessionMapping(
           sessionHash,
           selectedAccount.accountId,
-          selectedAccount.accountType
+          selectedAccount.accountType,
+          normalizedOauthProvider
         )
         logger.info(
           `🎯 Created new sticky session mapping: ${selectedAccount.name} (${selectedAccount.accountId}, ${selectedAccount.accountType}) for session ${sessionHash}`
@@ -189,7 +223,18 @@ class UnifiedGeminiScheduler {
   }
 
   // 📋 获取所有可用账户
-  async _getAllAvailableAccounts(apiKeyData, requestedModel = null, allowApiAccounts = false) {
+  async _getAllAvailableAccounts(
+    apiKeyData,
+    requestedModel = null,
+    allowApiAccountsOrOptions = false
+  ) {
+    const options =
+      allowApiAccountsOrOptions && typeof allowApiAccountsOrOptions === 'object'
+        ? allowApiAccountsOrOptions
+        : { allowApiAccounts: allowApiAccountsOrOptions }
+    const { allowApiAccounts = false, oauthProvider = null } = options
+    const normalizedOauthProvider = oauthProvider ? normalizeOauthProvider(oauthProvider) : null
+
     const availableAccounts = []
 
     // 如果API Key绑定了专属账户，优先返回
@@ -254,6 +299,12 @@ class UnifiedGeminiScheduler {
           this._isActive(boundAccount.isActive) &&
           boundAccount.status !== 'error'
         ) {
+          if (
+            normalizedOauthProvider &&
+            normalizeOauthProvider(boundAccount.oauthProvider) !== normalizedOauthProvider
+          ) {
+            return availableAccounts
+          }
           const isRateLimited = await this.isAccountRateLimited(boundAccount.id)
           if (!isRateLimited) {
             // 检查模型支持
@@ -303,6 +354,12 @@ class UnifiedGeminiScheduler {
         (account.accountType === 'shared' || !account.accountType) && // 兼容旧数据
         this._isSchedulable(account.schedulable)
       ) {
+        if (
+          normalizedOauthProvider &&
+          normalizeOauthProvider(account.oauthProvider) !== normalizedOauthProvider
+        ) {
+          continue
+        }
         // 检查是否可调度
 
         // 检查token是否过期
@@ -437,9 +494,10 @@ class UnifiedGeminiScheduler {
   }
 
   // 🔗 获取会话映射
-  async _getSessionMapping(sessionHash) {
+  async _getSessionMapping(sessionHash, oauthProvider = null) {
     const client = redis.getClientSafe()
-    const mappingData = await client.get(`${this.SESSION_MAPPING_PREFIX}${sessionHash}`)
+    const key = this._getSessionMappingKey(sessionHash, oauthProvider)
+    const mappingData = key ? await client.get(key) : null
 
     if (mappingData) {
       try {
@@ -454,27 +512,42 @@ class UnifiedGeminiScheduler {
   }
 
   // 💾 设置会话映射
-  async _setSessionMapping(sessionHash, accountId, accountType) {
+  async _setSessionMapping(sessionHash, accountId, accountType, oauthProvider = null) {
     const client = redis.getClientSafe()
     const mappingData = JSON.stringify({ accountId, accountType })
     // 依据配置设置TTL（小时）
     const appConfig = require('../../config/config')
     const ttlHours = appConfig.session?.stickyTtlHours || 1
     const ttlSeconds = Math.max(1, Math.floor(ttlHours * 60 * 60))
-    await client.setex(`${this.SESSION_MAPPING_PREFIX}${sessionHash}`, ttlSeconds, mappingData)
+    const key = this._getSessionMappingKey(sessionHash, oauthProvider)
+    if (!key) {
+      return
+    }
+    await client.setex(key, ttlSeconds, mappingData)
   }
 
   // 🗑️ 删除会话映射
   async _deleteSessionMapping(sessionHash) {
     const client = redis.getClientSafe()
-    await client.del(`${this.SESSION_MAPPING_PREFIX}${sessionHash}`)
+    if (!sessionHash) {
+      return
+    }
+
+    const keys = [this._getSessionMappingKey(sessionHash)]
+    for (const provider of KNOWN_OAUTH_PROVIDERS) {
+      keys.push(this._getSessionMappingKey(sessionHash, provider))
+    }
+    await client.del(keys.filter(Boolean))
   }
 
   // 🔁 续期统一调度会话映射TTL（针对 unified_gemini_session_mapping:* 键），遵循会话配置
-  async _extendSessionMappingTTL(sessionHash) {
+  async _extendSessionMappingTTL(sessionHash, oauthProvider = null) {
     try {
       const client = redis.getClientSafe()
-      const key = `${this.SESSION_MAPPING_PREFIX}${sessionHash}`
+      const key = this._getSessionMappingKey(sessionHash, oauthProvider)
+      if (!key) {
+        return false
+      }
       const remainingTTL = await client.ttl(key)
 
       if (remainingTTL === -2) {

src/services/unifiedOpenAIScheduler.js

@@ -9,6 +9,26 @@ class UnifiedOpenAIScheduler {
     this.SESSION_MAPPING_PREFIX = 'unified_openai_session_mapping:'
   }
 
+  // 🔢 按优先级和最后使用时间排序账户（与 Claude/Gemini 调度保持一致）
+  _sortAccountsByPriority(accounts) {
+    return accounts.sort((a, b) => {
+      const aPriority = Number.parseInt(a.priority, 10)
+      const bPriority = Number.parseInt(b.priority, 10)
+      const normalizedAPriority = Number.isFinite(aPriority) ? aPriority : 50
+      const normalizedBPriority = Number.isFinite(bPriority) ? bPriority : 50
+
+      // 首先按优先级排序（数字越小优先级越高）
+      if (normalizedAPriority !== normalizedBPriority) {
+        return normalizedAPriority - normalizedBPriority
+      }
+
+      // 优先级相同时，按最后使用时间排序（最久未使用的优先）
+      const aLastUsed = new Date(a.lastUsedAt || 0).getTime()
+      const bLastUsed = new Date(b.lastUsedAt || 0).getTime()
+      return aLastUsed - bLastUsed
+    })
+  }
+
   // 🔧 辅助方法：检查账户是否可调度（兼容字符串和布尔值）
   _isSchedulable(schedulable) {
     // 如果是 undefined 或 null，默认为可调度
@@ -244,13 +264,7 @@ class UnifiedOpenAIScheduler {
             `🎯 Using bound dedicated ${accountType} account: ${boundAccount.name} (${boundAccount.id}) for API key ${apiKeyData.name}`
           )
           // 更新账户的最后使用时间
-          if (accountType === 'openai') {
-            await openaiAccountService.recordUsage(boundAccount.id, 0)
-          } else {
-            await openaiResponsesAccountService.updateAccount(boundAccount.id, {
-              lastUsedAt: new Date().toISOString()
-            })
-          }
+          await this.updateAccountLastUsed(boundAccount.id, accountType)
           return {
             accountId: boundAccount.id,
             accountType
@@ -292,7 +306,7 @@ class UnifiedOpenAIScheduler {
               `🎯 Using sticky session account: ${mappedAccount.accountId} (${mappedAccount.accountType}) for session ${sessionHash}`
             )
             // 更新账户的最后使用时间
-            await openaiAccountService.recordUsage(mappedAccount.accountId, 0)
+            await this.updateAccountLastUsed(mappedAccount.accountId, mappedAccount.accountType)
             return mappedAccount
           } else {
             logger.warn(
@@ -321,12 +335,8 @@ class UnifiedOpenAIScheduler {
         }
       }
 
-      // 按最后使用时间排序（最久未使用的优先，与 Claude 保持一致）
-      const sortedAccounts = availableAccounts.sort((a, b) => {
-        const aLastUsed = new Date(a.lastUsedAt || 0).getTime()
-        const bLastUsed = new Date(b.lastUsedAt || 0).getTime()
-        return aLastUsed - bLastUsed // 最久未使用的优先
-      })
+      // 按优先级和最后使用时间排序（与 Claude/Gemini 调度保持一致）
+      const sortedAccounts = this._sortAccountsByPriority(availableAccounts)
 
       // 选择第一个账户
       const selectedAccount = sortedAccounts[0]
@@ -344,11 +354,11 @@ class UnifiedOpenAIScheduler {
       }
 
       logger.info(
-        `🎯 Selected account: ${selectedAccount.name} (${selectedAccount.accountId}, ${selectedAccount.accountType}) for API key ${apiKeyData.name}`
+        `🎯 Selected account: ${selectedAccount.name} (${selectedAccount.accountId}, ${selectedAccount.accountType}, priority: ${selectedAccount.priority || 50}) for API key ${apiKeyData.name}`
       )
 
       // 更新账户的最后使用时间
-      await openaiAccountService.recordUsage(selectedAccount.accountId, 0)
+      await this.updateAccountLastUsed(selectedAccount.accountId, selectedAccount.accountType)
 
       return {
         accountId: selectedAccount.accountId,
@@ -494,21 +504,6 @@ class UnifiedOpenAIScheduler {
     return availableAccounts
   }
 
-  // 🔢 按优先级和最后使用时间排序账户（已废弃，改为与 Claude 保持一致，只按最后使用时间排序）
-  // _sortAccountsByPriority(accounts) {
-  //   return accounts.sort((a, b) => {
-  //     // 首先按优先级排序（数字越小优先级越高）
-  //     if (a.priority !== b.priority) {
-  //       return a.priority - b.priority
-  //     }
-
-  //     // 优先级相同时，按最后使用时间排序（最久未使用的优先）
-  //     const aLastUsed = new Date(a.lastUsedAt || 0).getTime()
-  //     const bLastUsed = new Date(b.lastUsedAt || 0).getTime()
-  //     return aLastUsed - bLastUsed
-  //   })
-  // }
-
   // 🔍 检查账户是否可用
   async _isAccountAvailable(accountId, accountType) {
     try {
@@ -817,7 +812,7 @@ class UnifiedOpenAIScheduler {
                 `🎯 Using sticky session account from group: ${mappedAccount.accountId} (${mappedAccount.accountType})`
               )
               // 更新账户的最后使用时间
-              await openaiAccountService.recordUsage(mappedAccount.accountId, 0)
+              await this.updateAccountLastUsed(mappedAccount.accountId, mappedAccount.accountType)
               return mappedAccount
             }
           }
@@ -909,12 +904,8 @@ class UnifiedOpenAIScheduler {
         throw error
       }
 
-      // 按最后使用时间排序（最久未使用的优先，与 Claude 保持一致）
-      const sortedAccounts = availableAccounts.sort((a, b) => {
-        const aLastUsed = new Date(a.lastUsedAt || 0).getTime()
-        const bLastUsed = new Date(b.lastUsedAt || 0).getTime()
-        return aLastUsed - bLastUsed // 最久未使用的优先
-      })
+      // 按优先级和最后使用时间排序（与 Claude/Gemini 调度保持一致）
+      const sortedAccounts = this._sortAccountsByPriority(availableAccounts)
 
       // 选择第一个账户
       const selectedAccount = sortedAccounts[0]
@@ -932,11 +923,11 @@ class UnifiedOpenAIScheduler {
       }
 
       logger.info(
-        `🎯 Selected account from group: ${selectedAccount.name} (${selectedAccount.accountId})`
+        `🎯 Selected account from group: ${selectedAccount.name} (${selectedAccount.accountId}, ${selectedAccount.accountType}, priority: ${selectedAccount.priority || 50})`
       )
 
       // 更新账户的最后使用时间
-      await openaiAccountService.recordUsage(selectedAccount.accountId, 0)
+      await this.updateAccountLastUsed(selectedAccount.accountId, selectedAccount.accountType)
 
       return {
         accountId: selectedAccount.accountId,
@@ -958,9 +949,12 @@ class UnifiedOpenAIScheduler {
   async updateAccountLastUsed(accountId, accountType) {
     try {
       if (accountType === 'openai') {
-        await openaiAccountService.updateAccount(accountId, {
-          lastUsedAt: new Date().toISOString()
-        })
+        await openaiAccountService.recordUsage(accountId, 0)
+        return
+      }
+
+      if (accountType === 'openai-responses') {
+        await openaiResponsesAccountService.recordUsage(accountId, 0)
       }
     } catch (error) {
       logger.warn(`⚠️ Failed to update last used time for account ${accountId}:`, error)

src/services/userMessageQueueService.js

@@ -0,0 +1,359 @@
+/**
+ * 用户消息队列服务
+ * 为 Claude 账户实现基于消息类型的串行排队机制
+ *
+ * 当请求的最后一条消息是用户输入（role: user）时，
+ * 同一账户的此类请求需要串行等待，并在请求之间添加延迟
+ */
+
+const { v4: uuidv4 } = require('uuid')
+const redis = require('../models/redis')
+const config = require('../../config/config')
+const logger = require('../utils/logger')
+
+// 清理任务间隔
+const CLEANUP_INTERVAL_MS = 60000 // 1分钟
+
+// 轮询等待配置
+const POLL_INTERVAL_BASE_MS = 50 // 基础轮询间隔
+const POLL_INTERVAL_MAX_MS = 500 // 最大轮询间隔
+const POLL_BACKOFF_FACTOR = 1.5 // 退避因子
+
+class UserMessageQueueService {
+  constructor() {
+    this.cleanupTimer = null
+  }
+
+  /**
+   * 检测请求是否为真正的用户消息请求
+   * 区分真正的用户输入和 tool_result 消息
+   *
+   * Claude API 消息格式：
+   * - 用户文本消息: { role: 'user', content: 'text' } 或 { role: 'user', content: [{ type: 'text', text: '...' }] }
+   * - 工具结果消息: { role: 'user', content: [{ type: 'tool_result', tool_use_id: '...', content: '...' }] }
+   *
+   * @param {Object} requestBody - 请求体
+   * @returns {boolean} - 是否为真正的用户消息（排除 tool_result）
+   */
+  isUserMessageRequest(requestBody) {
+    const messages = requestBody?.messages
+    if (!Array.isArray(messages) || messages.length === 0) {
+      return false
+    }
+    const lastMessage = messages[messages.length - 1]
+
+    // 检查 role 是否为 user
+    if (lastMessage?.role !== 'user') {
+      return false
+    }
+
+    // 检查 content 是否包含 tool_result 类型
+    const { content } = lastMessage
+    if (Array.isArray(content)) {
+      // 如果 content 数组中任何元素是 tool_result，则不是真正的用户消息
+      const hasToolResult = content.some(
+        (block) => block?.type === 'tool_result' || block?.type === 'tool_use_result'
+      )
+      if (hasToolResult) {
+        return false
+      }
+    }
+
+    // role 是 user 且不包含 tool_result，是真正的用户消息
+    return true
+  }
+
+  /**
+   * 获取当前配置（支持 Web 界面配置优先）
+   * @returns {Promise<Object>} 配置对象
+   */
+  async getConfig() {
+    // 默认配置（防止 config.userMessageQueue 未定义）
+    // 注意：优化后的默认值 - 锁持有时间从分钟级降到毫秒级，无需长等待
+    const queueConfig = config.userMessageQueue || {}
+    const defaults = {
+      enabled: queueConfig.enabled ?? false,
+      delayMs: queueConfig.delayMs ?? 200,
+      timeoutMs: queueConfig.timeoutMs ?? 5000, // 从 60000 降到 5000，因为锁持有时间短
+      lockTtlMs: queueConfig.lockTtlMs ?? 5000 // 从 120000 降到 5000，5秒足以覆盖请求发送
+    }
+
+    // 尝试从 claudeRelayConfigService 获取 Web 界面配置
+    try {
+      const claudeRelayConfigService = require('./claudeRelayConfigService')
+      const webConfig = await claudeRelayConfigService.getConfig()
+
+      return {
+        enabled:
+          webConfig.userMessageQueueEnabled !== undefined
+            ? webConfig.userMessageQueueEnabled
+            : defaults.enabled,
+        delayMs:
+          webConfig.userMessageQueueDelayMs !== undefined
+            ? webConfig.userMessageQueueDelayMs
+            : defaults.delayMs,
+        timeoutMs:
+          webConfig.userMessageQueueTimeoutMs !== undefined
+            ? webConfig.userMessageQueueTimeoutMs
+            : defaults.timeoutMs,
+        lockTtlMs:
+          webConfig.userMessageQueueLockTtlMs !== undefined
+            ? webConfig.userMessageQueueLockTtlMs
+            : defaults.lockTtlMs
+      }
+    } catch {
+      // 回退到环境变量配置
+      return defaults
+    }
+  }
+
+  /**
+   * 检查功能是否启用
+   * @returns {Promise<boolean>}
+   */
+  async isEnabled() {
+    const cfg = await this.getConfig()
+    return cfg.enabled === true
+  }
+
+  /**
+   * 获取账户队列锁（阻塞等待）
+   * @param {string} accountId - 账户ID
+   * @param {string} requestId - 请求ID（可选，会自动生成）
+   * @param {number} timeoutMs - 超时时间（可选，使用配置默认值）
+   * @param {Object} accountConfig - 账户级配置（可选），优先级高于全局配置
+   * @param {number} accountConfig.maxConcurrency - 账户级串行队列开关：>0启用，=0使用全局配置
+   * @returns {Promise<{acquired: boolean, requestId: string, error?: string}>}
+   */
+  async acquireQueueLock(accountId, requestId = null, timeoutMs = null, accountConfig = null) {
+    const cfg = await this.getConfig()
+
+    // 账户级配置优先：maxConcurrency > 0 时强制启用，忽略全局开关
+    let queueEnabled = cfg.enabled
+    if (accountConfig && accountConfig.maxConcurrency > 0) {
+      queueEnabled = true
+      logger.debug(
+        `📬 User message queue: account-level queue enabled for account ${accountId} (maxConcurrency=${accountConfig.maxConcurrency})`
+      )
+    }
+
+    if (!queueEnabled) {
+      return { acquired: true, requestId: requestId || uuidv4(), skipped: true }
+    }
+
+    const reqId = requestId || uuidv4()
+    const timeout = timeoutMs || cfg.timeoutMs
+    const startTime = Date.now()
+    let retryCount = 0
+
+    logger.debug(`📬 User message queue: attempting to acquire lock for account ${accountId}`, {
+      requestId: reqId,
+      timeoutMs: timeout
+    })
+
+    while (Date.now() - startTime < timeout) {
+      const result = await redis.acquireUserMessageLock(
+        accountId,
+        reqId,
+        cfg.lockTtlMs,
+        cfg.delayMs
+      )
+
+      // 检测 Redis 错误，立即返回系统错误而非继续轮询
+      if (result.redisError) {
+        logger.error(`📬 User message queue: Redis error while acquiring lock`, {
+          accountId,
+          requestId: reqId,
+          errorMessage: result.errorMessage
+        })
+        return {
+          acquired: false,
+          requestId: reqId,
+          error: 'queue_backend_error',
+          errorMessage: result.errorMessage
+        }
+      }
+
+      if (result.acquired) {
+        logger.debug(`📬 User message queue: lock acquired for account ${accountId}`, {
+          requestId: reqId,
+          waitedMs: Date.now() - startTime,
+          retries: retryCount
+        })
+        return { acquired: true, requestId: reqId }
+      }
+
+      // 需要等待
+      if (result.waitMs > 0) {
+        // 需要延迟（上一个请求刚完成）
+        await this._sleep(Math.min(result.waitMs, timeout - (Date.now() - startTime)))
+      } else {
+        // 锁被占用，使用指数退避轮询等待
+        const basePollInterval = Math.min(
+          POLL_INTERVAL_BASE_MS * Math.pow(POLL_BACKOFF_FACTOR, retryCount),
+          POLL_INTERVAL_MAX_MS
+        )
+        // 添加 ±15% 随机抖动，避免高并发下的周期性碰撞
+        const jitter = basePollInterval * (0.85 + Math.random() * 0.3)
+        const pollInterval = Math.min(jitter, POLL_INTERVAL_MAX_MS)
+        await this._sleep(pollInterval)
+        retryCount++
+      }
+    }
+
+    // 超时
+    logger.warn(`📬 User message queue: timeout waiting for lock`, {
+      accountId,
+      requestId: reqId,
+      timeoutMs: timeout
+    })
+
+    return {
+      acquired: false,
+      requestId: reqId,
+      error: 'queue_timeout'
+    }
+  }
+
+  /**
+   * 释放账户队列锁
+   * @param {string} accountId - 账户ID
+   * @param {string} requestId - 请求ID
+   * @returns {Promise<boolean>}
+   */
+  async releaseQueueLock(accountId, requestId) {
+    if (!accountId || !requestId) {
+      return false
+    }
+
+    const released = await redis.releaseUserMessageLock(accountId, requestId)
+
+    if (released) {
+      logger.debug(`📬 User message queue: lock released for account ${accountId}`, {
+        requestId
+      })
+    } else {
+      logger.warn(`📬 User message queue: failed to release lock (not owner?)`, {
+        accountId,
+        requestId
+      })
+    }
+
+    return released
+  }
+
+  /**
+   * 获取队列统计信息
+   * @param {string} accountId - 账户ID
+   * @returns {Promise<Object>}
+   */
+  async getQueueStats(accountId) {
+    return await redis.getUserMessageQueueStats(accountId)
+  }
+
+  /**
+   * 服务启动时清理所有残留的队列锁
+   * 防止服务重启后旧锁阻塞新请求
+   * @returns {Promise<number>} 清理的锁数量
+   */
+  async cleanupStaleLocks() {
+    try {
+      const accountIds = await redis.scanUserMessageQueueLocks()
+      let cleanedCount = 0
+
+      for (const accountId of accountIds) {
+        try {
+          await redis.forceReleaseUserMessageLock(accountId)
+          cleanedCount++
+          logger.debug(`📬 User message queue: cleaned stale lock for account ${accountId}`)
+        } catch (error) {
+          logger.error(
+            `📬 User message queue: failed to clean lock for account ${accountId}:`,
+            error
+          )
+        }
+      }
+
+      if (cleanedCount > 0) {
+        logger.info(`📬 User message queue: cleaned ${cleanedCount} stale lock(s) on startup`)
+      }
+
+      return cleanedCount
+    } catch (error) {
+      logger.error('📬 User message queue: failed to cleanup stale locks on startup:', error)
+      return 0
+    }
+  }
+
+  /**
+   * 启动定时清理任务
+   * 始终启动，每次执行时检查配置以支持运行时动态启用/禁用
+   */
+  startCleanupTask() {
+    if (this.cleanupTimer) {
+      return
+    }
+
+    this.cleanupTimer = setInterval(async () => {
+      // 每次运行时检查配置，以便在运行时动态启用/禁用
+      const currentConfig = await this.getConfig()
+      if (!currentConfig.enabled) {
+        logger.debug('📬 User message queue: cleanup skipped (feature disabled)')
+        return
+      }
+      await this._cleanupOrphanLocks()
+    }, CLEANUP_INTERVAL_MS)
+
+    logger.info('📬 User message queue: cleanup task started')
+  }
+
+  /**
+   * 停止定时清理任务
+   */
+  stopCleanupTask() {
+    if (this.cleanupTimer) {
+      clearInterval(this.cleanupTimer)
+      this.cleanupTimer = null
+      logger.info('📬 User message queue: cleanup task stopped')
+    }
+  }
+
+  /**
+   * 清理孤儿锁
+   * 检测异常情况：锁存在但没有设置过期时间（lockTtlRaw === -1）
+   * 正常情况下所有锁都应该有 TTL，Redis 会自动过期
+   * @private
+   */
+  async _cleanupOrphanLocks() {
+    try {
+      const accountIds = await redis.scanUserMessageQueueLocks()
+
+      for (const accountId of accountIds) {
+        const stats = await redis.getUserMessageQueueStats(accountId)
+
+        // 检测异常情况：锁存在（isLocked=true）但没有过期时间（lockTtlRaw=-1）
+        // 正常创建的锁都带有 PX 过期时间，如果没有说明是异常状态
+        if (stats.isLocked && stats.lockTtlRaw === -1) {
+          logger.warn(
+            `📬 User message queue: cleaning up orphan lock without TTL for account ${accountId}`,
+            { lockHolder: stats.lockHolder }
+          )
+          await redis.forceReleaseUserMessageLock(accountId)
+        }
+      }
+    } catch (error) {
+      logger.error('📬 User message queue: cleanup task error:', error)
+    }
+  }
+
+  /**
+   * 睡眠辅助函数
+   * @param {number} ms - 毫秒
+   * @private
+   */
+  _sleep(ms) {
+    return new Promise((resolve) => setTimeout(resolve, ms))
+  }
+}
+
+module.exports = new UserMessageQueueService()

src/utils/anthropicRequestDump.js

@@ -0,0 +1,126 @@
+const path = require('path')
+const logger = require('./logger')
+const { getProjectRoot } = require('./projectPaths')
+const { safeRotatingAppend } = require('./safeRotatingAppend')
+
+const REQUEST_DUMP_ENV = 'ANTHROPIC_DEBUG_REQUEST_DUMP'
+const REQUEST_DUMP_MAX_BYTES_ENV = 'ANTHROPIC_DEBUG_REQUEST_DUMP_MAX_BYTES'
+const REQUEST_DUMP_FILENAME = 'anthropic-requests-dump.jsonl'
+
+function isEnabled() {
+  const raw = process.env[REQUEST_DUMP_ENV]
+  if (!raw) {
+    return false
+  }
+  return raw === '1' || raw.toLowerCase() === 'true'
+}
+
+function getMaxBytes() {
+  const raw = process.env[REQUEST_DUMP_MAX_BYTES_ENV]
+  if (!raw) {
+    return 2 * 1024 * 1024
+  }
+  const parsed = Number.parseInt(raw, 10)
+  if (!Number.isFinite(parsed) || parsed <= 0) {
+    return 2 * 1024 * 1024
+  }
+  return parsed
+}
+
+function maskSecret(value) {
+  if (value === null || value === undefined) {
+    return value
+  }
+  const str = String(value)
+  if (str.length <= 8) {
+    return '***'
+  }
+  return `${str.slice(0, 4)}...${str.slice(-4)}`
+}
+
+function sanitizeHeaders(headers) {
+  const sensitive = new Set([
+    'authorization',
+    'proxy-authorization',
+    'x-api-key',
+    'cookie',
+    'set-cookie',
+    'x-forwarded-for',
+    'x-real-ip'
+  ])
+
+  const out = {}
+  for (const [k, v] of Object.entries(headers || {})) {
+    const key = k.toLowerCase()
+    if (sensitive.has(key)) {
+      out[key] = maskSecret(v)
+      continue
+    }
+    out[key] = v
+  }
+  return out
+}
+
+function safeJsonStringify(payload, maxBytes) {
+  let json = ''
+  try {
+    json = JSON.stringify(payload)
+  } catch (e) {
+    return JSON.stringify({
+      type: 'anthropic_request_dump_error',
+      error: 'JSON.stringify_failed',
+      message: e?.message || String(e)
+    })
+  }
+
+  if (Buffer.byteLength(json, 'utf8') <= maxBytes) {
+    return json
+  }
+
+  const truncated = Buffer.from(json, 'utf8').subarray(0, maxBytes).toString('utf8')
+  return JSON.stringify({
+    type: 'anthropic_request_dump_truncated',
+    maxBytes,
+    originalBytes: Buffer.byteLength(json, 'utf8'),
+    partialJson: truncated
+  })
+}
+
+async function dumpAnthropicMessagesRequest(req, meta = {}) {
+  if (!isEnabled()) {
+    return
+  }
+
+  const maxBytes = getMaxBytes()
+  const filename = path.join(getProjectRoot(), REQUEST_DUMP_FILENAME)
+
+  const record = {
+    ts: new Date().toISOString(),
+    requestId: req?.requestId || null,
+    method: req?.method || null,
+    url: req?.originalUrl || req?.url || null,
+    ip: req?.ip || null,
+    meta,
+    headers: sanitizeHeaders(req?.headers || {}),
+    body: req?.body || null
+  }
+
+  const line = `${safeJsonStringify(record, maxBytes)}\n`
+
+  try {
+    await safeRotatingAppend(filename, line)
+  } catch (e) {
+    logger.warn('Failed to dump Anthropic request', {
+      filename,
+      requestId: req?.requestId || null,
+      error: e?.message || String(e)
+    })
+  }
+}
+
+module.exports = {
+  dumpAnthropicMessagesRequest,
+  REQUEST_DUMP_ENV,
+  REQUEST_DUMP_MAX_BYTES_ENV,
+  REQUEST_DUMP_FILENAME
+}

src/utils/anthropicResponseDump.js

@@ -0,0 +1,125 @@
+const path = require('path')
+const logger = require('./logger')
+const { getProjectRoot } = require('./projectPaths')
+const { safeRotatingAppend } = require('./safeRotatingAppend')
+
+const RESPONSE_DUMP_ENV = 'ANTHROPIC_DEBUG_RESPONSE_DUMP'
+const RESPONSE_DUMP_MAX_BYTES_ENV = 'ANTHROPIC_DEBUG_RESPONSE_DUMP_MAX_BYTES'
+const RESPONSE_DUMP_FILENAME = 'anthropic-responses-dump.jsonl'
+
+function isEnabled() {
+  const raw = process.env[RESPONSE_DUMP_ENV]
+  if (!raw) {
+    return false
+  }
+  return raw === '1' || raw.toLowerCase() === 'true'
+}
+
+function getMaxBytes() {
+  const raw = process.env[RESPONSE_DUMP_MAX_BYTES_ENV]
+  if (!raw) {
+    return 2 * 1024 * 1024
+  }
+  const parsed = Number.parseInt(raw, 10)
+  if (!Number.isFinite(parsed) || parsed <= 0) {
+    return 2 * 1024 * 1024
+  }
+  return parsed
+}
+
+function safeJsonStringify(payload, maxBytes) {
+  let json = ''
+  try {
+    json = JSON.stringify(payload)
+  } catch (e) {
+    return JSON.stringify({
+      type: 'anthropic_response_dump_error',
+      error: 'JSON.stringify_failed',
+      message: e?.message || String(e)
+    })
+  }
+
+  if (Buffer.byteLength(json, 'utf8') <= maxBytes) {
+    return json
+  }
+
+  const truncated = Buffer.from(json, 'utf8').subarray(0, maxBytes).toString('utf8')
+  return JSON.stringify({
+    type: 'anthropic_response_dump_truncated',
+    maxBytes,
+    originalBytes: Buffer.byteLength(json, 'utf8'),
+    partialJson: truncated
+  })
+}
+
+function summarizeAnthropicResponseBody(body) {
+  const content = Array.isArray(body?.content) ? body.content : []
+  const toolUses = content.filter((b) => b && b.type === 'tool_use')
+  const texts = content
+    .filter((b) => b && b.type === 'text' && typeof b.text === 'string')
+    .map((b) => b.text)
+    .join('')
+
+  return {
+    id: body?.id || null,
+    model: body?.model || null,
+    stop_reason: body?.stop_reason || null,
+    usage: body?.usage || null,
+    content_blocks: content.map((b) => (b ? b.type : null)).filter(Boolean),
+    tool_use_names: toolUses.map((b) => b.name).filter(Boolean),
+    text_preview: texts ? texts.slice(0, 800) : ''
+  }
+}
+
+async function dumpAnthropicResponse(req, responseInfo, meta = {}) {
+  if (!isEnabled()) {
+    return
+  }
+
+  const maxBytes = getMaxBytes()
+  const filename = path.join(getProjectRoot(), RESPONSE_DUMP_FILENAME)
+
+  const record = {
+    ts: new Date().toISOString(),
+    requestId: req?.requestId || null,
+    url: req?.originalUrl || req?.url || null,
+    meta,
+    response: responseInfo
+  }
+
+  const line = `${safeJsonStringify(record, maxBytes)}\n`
+  try {
+    await safeRotatingAppend(filename, line)
+  } catch (e) {
+    logger.warn('Failed to dump Anthropic response', {
+      filename,
+      requestId: req?.requestId || null,
+      error: e?.message || String(e)
+    })
+  }
+}
+
+async function dumpAnthropicNonStreamResponse(req, statusCode, body, meta = {}) {
+  return dumpAnthropicResponse(
+    req,
+    { kind: 'non-stream', statusCode, summary: summarizeAnthropicResponseBody(body), body },
+    meta
+  )
+}
+
+async function dumpAnthropicStreamSummary(req, summary, meta = {}) {
+  return dumpAnthropicResponse(req, { kind: 'stream', summary }, meta)
+}
+
+async function dumpAnthropicStreamError(req, error, meta = {}) {
+  return dumpAnthropicResponse(req, { kind: 'stream-error', error }, meta)
+}
+
+module.exports = {
+  dumpAnthropicNonStreamResponse,
+  dumpAnthropicStreamSummary,
+  dumpAnthropicStreamError,
+  RESPONSE_DUMP_ENV,
+  RESPONSE_DUMP_MAX_BYTES_ENV,
+  RESPONSE_DUMP_FILENAME
+}

src/utils/antigravityModel.js

@@ -0,0 +1,138 @@
+const DEFAULT_ANTIGRAVITY_MODEL = 'gemini-2.5-flash'
+
+const UPSTREAM_TO_ALIAS = {
+  'rev19-uic3-1p': 'gemini-2.5-computer-use-preview-10-2025',
+  'gemini-3-pro-image': 'gemini-3-pro-image-preview',
+  'gemini-3-pro-high': 'gemini-3-pro-preview',
+  'gemini-3-flash': 'gemini-3-flash-preview',
+  'claude-sonnet-4-5': 'gemini-claude-sonnet-4-5',
+  'claude-sonnet-4-5-thinking': 'gemini-claude-sonnet-4-5-thinking',
+  'claude-opus-4-5-thinking': 'gemini-claude-opus-4-5-thinking',
+  chat_20706: '',
+  chat_23310: '',
+  'gemini-2.5-flash-thinking': '',
+  'gemini-3-pro-low': '',
+  'gemini-2.5-pro': ''
+}
+
+const ALIAS_TO_UPSTREAM = {
+  'gemini-2.5-computer-use-preview-10-2025': 'rev19-uic3-1p',
+  'gemini-3-pro-image-preview': 'gemini-3-pro-image',
+  'gemini-3-pro-preview': 'gemini-3-pro-high',
+  'gemini-3-flash-preview': 'gemini-3-flash',
+  'gemini-claude-sonnet-4-5': 'claude-sonnet-4-5',
+  'gemini-claude-sonnet-4-5-thinking': 'claude-sonnet-4-5-thinking',
+  'gemini-claude-opus-4-5-thinking': 'claude-opus-4-5-thinking'
+}
+
+const ANTIGRAVITY_MODEL_METADATA = {
+  'gemini-2.5-flash': {
+    thinking: { min: 0, max: 24576, zeroAllowed: true, dynamicAllowed: true },
+    name: 'models/gemini-2.5-flash'
+  },
+  'gemini-2.5-flash-lite': {
+    thinking: { min: 0, max: 24576, zeroAllowed: true, dynamicAllowed: true },
+    name: 'models/gemini-2.5-flash-lite'
+  },
+  'gemini-2.5-computer-use-preview-10-2025': {
+    name: 'models/gemini-2.5-computer-use-preview-10-2025'
+  },
+  'gemini-3-pro-preview': {
+    thinking: {
+      min: 128,
+      max: 32768,
+      zeroAllowed: false,
+      dynamicAllowed: true,
+      levels: ['low', 'high']
+    },
+    name: 'models/gemini-3-pro-preview'
+  },
+  'gemini-3-pro-image-preview': {
+    thinking: {
+      min: 128,
+      max: 32768,
+      zeroAllowed: false,
+      dynamicAllowed: true,
+      levels: ['low', 'high']
+    },
+    name: 'models/gemini-3-pro-image-preview'
+  },
+  'gemini-3-flash-preview': {
+    thinking: {
+      min: 128,
+      max: 32768,
+      zeroAllowed: false,
+      dynamicAllowed: true,
+      levels: ['minimal', 'low', 'medium', 'high']
+    },
+    name: 'models/gemini-3-flash-preview'
+  },
+  'gemini-claude-sonnet-4-5-thinking': {
+    thinking: { min: 1024, max: 200000, zeroAllowed: false, dynamicAllowed: true },
+    maxCompletionTokens: 64000
+  },
+  'gemini-claude-opus-4-5-thinking': {
+    thinking: { min: 1024, max: 200000, zeroAllowed: false, dynamicAllowed: true },
+    maxCompletionTokens: 64000
+  }
+}
+
+function normalizeAntigravityModelInput(model, defaultModel = DEFAULT_ANTIGRAVITY_MODEL) {
+  if (!model) {
+    return defaultModel
+  }
+  return model.startsWith('models/') ? model.slice('models/'.length) : model
+}
+
+function getAntigravityModelAlias(modelName) {
+  const normalized = normalizeAntigravityModelInput(modelName)
+  if (Object.prototype.hasOwnProperty.call(UPSTREAM_TO_ALIAS, normalized)) {
+    return UPSTREAM_TO_ALIAS[normalized]
+  }
+  return normalized
+}
+
+function getAntigravityModelMetadata(modelName) {
+  const normalized = normalizeAntigravityModelInput(modelName)
+  if (Object.prototype.hasOwnProperty.call(ANTIGRAVITY_MODEL_METADATA, normalized)) {
+    return ANTIGRAVITY_MODEL_METADATA[normalized]
+  }
+  if (normalized.startsWith('claude-')) {
+    const prefixed = `gemini-${normalized}`
+    if (Object.prototype.hasOwnProperty.call(ANTIGRAVITY_MODEL_METADATA, prefixed)) {
+      return ANTIGRAVITY_MODEL_METADATA[prefixed]
+    }
+    const thinkingAlias = `${prefixed}-thinking`
+    if (Object.prototype.hasOwnProperty.call(ANTIGRAVITY_MODEL_METADATA, thinkingAlias)) {
+      return ANTIGRAVITY_MODEL_METADATA[thinkingAlias]
+    }
+  }
+  return null
+}
+
+function mapAntigravityUpstreamModel(model) {
+  const normalized = normalizeAntigravityModelInput(model)
+  let upstream = Object.prototype.hasOwnProperty.call(ALIAS_TO_UPSTREAM, normalized)
+    ? ALIAS_TO_UPSTREAM[normalized]
+    : normalized
+
+  if (upstream.startsWith('gemini-claude-')) {
+    upstream = upstream.replace(/^gemini-/, '')
+  }
+
+  const mapping = {
+    // Opus：上游更常见的是 thinking 变体（CLIProxyAPI 也按此处理）
+    'claude-opus-4-5': 'claude-opus-4-5-thinking',
+    // Gemini thinking 变体回退
+    'gemini-2.5-flash-thinking': 'gemini-2.5-flash'
+  }
+
+  return mapping[upstream] || upstream
+}
+
+module.exports = {
+  normalizeAntigravityModelInput,
+  getAntigravityModelAlias,
+  getAntigravityModelMetadata,
+  mapAntigravityUpstreamModel
+}

src/utils/antigravityUpstreamDump.js

@@ -0,0 +1,121 @@
+const path = require('path')
+const logger = require('./logger')
+const { getProjectRoot } = require('./projectPaths')
+const { safeRotatingAppend } = require('./safeRotatingAppend')
+
+const UPSTREAM_REQUEST_DUMP_ENV = 'ANTIGRAVITY_DEBUG_UPSTREAM_REQUEST_DUMP'
+const UPSTREAM_REQUEST_DUMP_MAX_BYTES_ENV = 'ANTIGRAVITY_DEBUG_UPSTREAM_REQUEST_DUMP_MAX_BYTES'
+const UPSTREAM_REQUEST_DUMP_FILENAME = 'antigravity-upstream-requests-dump.jsonl'
+
+function isEnabled() {
+  const raw = process.env[UPSTREAM_REQUEST_DUMP_ENV]
+  if (!raw) {
+    return false
+  }
+  const normalized = String(raw).trim().toLowerCase()
+  return normalized === '1' || normalized === 'true'
+}
+
+function getMaxBytes() {
+  const raw = process.env[UPSTREAM_REQUEST_DUMP_MAX_BYTES_ENV]
+  if (!raw) {
+    return 2 * 1024 * 1024
+  }
+  const parsed = Number.parseInt(raw, 10)
+  if (!Number.isFinite(parsed) || parsed <= 0) {
+    return 2 * 1024 * 1024
+  }
+  return parsed
+}
+
+function redact(value) {
+  if (!value) {
+    return value
+  }
+  const s = String(value)
+  if (s.length <= 10) {
+    return '***'
+  }
+  return `${s.slice(0, 3)}...${s.slice(-4)}`
+}
+
+function safeJsonStringify(payload, maxBytes) {
+  let json = ''
+  try {
+    json = JSON.stringify(payload)
+  } catch (e) {
+    return JSON.stringify({
+      type: 'antigravity_upstream_dump_error',
+      error: 'JSON.stringify_failed',
+      message: e?.message || String(e)
+    })
+  }
+
+  if (Buffer.byteLength(json, 'utf8') <= maxBytes) {
+    return json
+  }
+
+  const truncated = Buffer.from(json, 'utf8').subarray(0, maxBytes).toString('utf8')
+  return JSON.stringify({
+    type: 'antigravity_upstream_dump_truncated',
+    maxBytes,
+    originalBytes: Buffer.byteLength(json, 'utf8'),
+    partialJson: truncated
+  })
+}
+
+async function dumpAntigravityUpstreamRequest(requestInfo) {
+  if (!isEnabled()) {
+    return
+  }
+
+  const maxBytes = getMaxBytes()
+  const filename = path.join(getProjectRoot(), UPSTREAM_REQUEST_DUMP_FILENAME)
+
+  const record = {
+    ts: new Date().toISOString(),
+    type: 'antigravity_upstream_request',
+    requestId: requestInfo?.requestId || null,
+    model: requestInfo?.model || null,
+    stream: Boolean(requestInfo?.stream),
+    url: requestInfo?.url || null,
+    baseUrl: requestInfo?.baseUrl || null,
+    params: requestInfo?.params || null,
+    headers: requestInfo?.headers
+      ? {
+          Host: requestInfo.headers.Host || requestInfo.headers.host || null,
+          'User-Agent':
+            requestInfo.headers['User-Agent'] || requestInfo.headers['user-agent'] || null,
+          Authorization: (() => {
+            const raw = requestInfo.headers.Authorization || requestInfo.headers.authorization
+            if (!raw) {
+              return null
+            }
+            const value = String(raw)
+            const m = value.match(/^Bearer\\s+(.+)$/i)
+            const token = m ? m[1] : value
+            return `Bearer ${redact(token)}`
+          })()
+        }
+      : null,
+    envelope: requestInfo?.envelope || null
+  }
+
+  const line = `${safeJsonStringify(record, maxBytes)}\n`
+  try {
+    await safeRotatingAppend(filename, line)
+  } catch (e) {
+    logger.warn('Failed to dump Antigravity upstream request', {
+      filename,
+      requestId: requestInfo?.requestId || null,
+      error: e?.message || String(e)
+    })
+  }
+}
+
+module.exports = {
+  dumpAntigravityUpstreamRequest,
+  UPSTREAM_REQUEST_DUMP_ENV,
+  UPSTREAM_REQUEST_DUMP_MAX_BYTES_ENV,
+  UPSTREAM_REQUEST_DUMP_FILENAME
+}

src/utils/antigravityUpstreamResponseDump.js

@@ -0,0 +1,175 @@
+const path = require('path')
+const logger = require('./logger')
+const { getProjectRoot } = require('./projectPaths')
+const { safeRotatingAppend } = require('./safeRotatingAppend')
+
+const UPSTREAM_RESPONSE_DUMP_ENV = 'ANTIGRAVITY_DEBUG_UPSTREAM_RESPONSE_DUMP'
+const UPSTREAM_RESPONSE_DUMP_MAX_BYTES_ENV = 'ANTIGRAVITY_DEBUG_UPSTREAM_RESPONSE_DUMP_MAX_BYTES'
+const UPSTREAM_RESPONSE_DUMP_FILENAME = 'antigravity-upstream-responses-dump.jsonl'
+
+function isEnabled() {
+  const raw = process.env[UPSTREAM_RESPONSE_DUMP_ENV]
+  if (!raw) {
+    return false
+  }
+  const normalized = String(raw).trim().toLowerCase()
+  return normalized === '1' || normalized === 'true'
+}
+
+function getMaxBytes() {
+  const raw = process.env[UPSTREAM_RESPONSE_DUMP_MAX_BYTES_ENV]
+  if (!raw) {
+    return 2 * 1024 * 1024
+  }
+  const parsed = Number.parseInt(raw, 10)
+  if (!Number.isFinite(parsed) || parsed <= 0) {
+    return 2 * 1024 * 1024
+  }
+  return parsed
+}
+
+function safeJsonStringify(payload, maxBytes) {
+  let json = ''
+  try {
+    json = JSON.stringify(payload)
+  } catch (e) {
+    return JSON.stringify({
+      type: 'antigravity_upstream_response_dump_error',
+      error: 'JSON.stringify_failed',
+      message: e?.message || String(e)
+    })
+  }
+
+  if (Buffer.byteLength(json, 'utf8') <= maxBytes) {
+    return json
+  }
+
+  const truncated = Buffer.from(json, 'utf8').subarray(0, maxBytes).toString('utf8')
+  return JSON.stringify({
+    type: 'antigravity_upstream_response_dump_truncated',
+    maxBytes,
+    originalBytes: Buffer.byteLength(json, 'utf8'),
+    partialJson: truncated
+  })
+}
+
+/**
+ * 记录 Antigravity 上游 API 的响应
+ * @param {Object} responseInfo - 响应信息
+ * @param {string} responseInfo.requestId - 请求 ID
+ * @param {string} responseInfo.model - 模型名称
+ * @param {number} responseInfo.statusCode - HTTP 状态码
+ * @param {string} responseInfo.statusText - HTTP 状态文本
+ * @param {Object} responseInfo.headers - 响应头
+ * @param {string} responseInfo.responseType - 响应类型 (stream/non-stream/error)
+ * @param {Object} responseInfo.summary - 响应摘要
+ * @param {Object} responseInfo.error - 错误信息（如果有）
+ */
+async function dumpAntigravityUpstreamResponse(responseInfo) {
+  if (!isEnabled()) {
+    return
+  }
+
+  const maxBytes = getMaxBytes()
+  const filename = path.join(getProjectRoot(), UPSTREAM_RESPONSE_DUMP_FILENAME)
+
+  const record = {
+    ts: new Date().toISOString(),
+    type: 'antigravity_upstream_response',
+    requestId: responseInfo?.requestId || null,
+    model: responseInfo?.model || null,
+    statusCode: responseInfo?.statusCode || null,
+    statusText: responseInfo?.statusText || null,
+    responseType: responseInfo?.responseType || null,
+    headers: responseInfo?.headers || null,
+    summary: responseInfo?.summary || null,
+    error: responseInfo?.error || null,
+    rawData: responseInfo?.rawData || null
+  }
+
+  const line = `${safeJsonStringify(record, maxBytes)}\n`
+  try {
+    await safeRotatingAppend(filename, line)
+  } catch (e) {
+    logger.warn('Failed to dump Antigravity upstream response', {
+      filename,
+      requestId: responseInfo?.requestId || null,
+      error: e?.message || String(e)
+    })
+  }
+}
+
+/**
+ * 记录 SSE 流中的每个事件（用于详细调试）
+ */
+async function dumpAntigravityStreamEvent(eventInfo) {
+  if (!isEnabled()) {
+    return
+  }
+
+  const maxBytes = getMaxBytes()
+  const filename = path.join(getProjectRoot(), UPSTREAM_RESPONSE_DUMP_FILENAME)
+
+  const record = {
+    ts: new Date().toISOString(),
+    type: 'antigravity_stream_event',
+    requestId: eventInfo?.requestId || null,
+    eventIndex: eventInfo?.eventIndex || null,
+    eventType: eventInfo?.eventType || null,
+    data: eventInfo?.data || null
+  }
+
+  const line = `${safeJsonStringify(record, maxBytes)}\n`
+  try {
+    await safeRotatingAppend(filename, line)
+  } catch (e) {
+    // 静默处理，避免日志过多
+  }
+}
+
+/**
+ * 记录流式响应的最终摘要
+ */
+async function dumpAntigravityStreamSummary(summaryInfo) {
+  if (!isEnabled()) {
+    return
+  }
+
+  const maxBytes = getMaxBytes()
+  const filename = path.join(getProjectRoot(), UPSTREAM_RESPONSE_DUMP_FILENAME)
+
+  const record = {
+    ts: new Date().toISOString(),
+    type: 'antigravity_stream_summary',
+    requestId: summaryInfo?.requestId || null,
+    model: summaryInfo?.model || null,
+    totalEvents: summaryInfo?.totalEvents || 0,
+    finishReason: summaryInfo?.finishReason || null,
+    hasThinking: summaryInfo?.hasThinking || false,
+    hasToolCalls: summaryInfo?.hasToolCalls || false,
+    toolCallNames: summaryInfo?.toolCallNames || [],
+    usage: summaryInfo?.usage || null,
+    error: summaryInfo?.error || null,
+    textPreview: summaryInfo?.textPreview || null
+  }
+
+  const line = `${safeJsonStringify(record, maxBytes)}\n`
+  try {
+    await safeRotatingAppend(filename, line)
+  } catch (e) {
+    logger.warn('Failed to dump Antigravity stream summary', {
+      filename,
+      requestId: summaryInfo?.requestId || null,
+      error: e?.message || String(e)
+    })
+  }
+}
+
+module.exports = {
+  dumpAntigravityUpstreamResponse,
+  dumpAntigravityStreamEvent,
+  dumpAntigravityStreamSummary,
+  UPSTREAM_RESPONSE_DUMP_ENV,
+  UPSTREAM_RESPONSE_DUMP_MAX_BYTES_ENV,
+  UPSTREAM_RESPONSE_DUMP_FILENAME
+}

src/utils/contents.js

@@ -79,11 +79,21 @@ const PROMPT_DEFINITIONS = {
     title: 'Claude Code Compact System Prompt Agent SDK2',
     text: "You are Claude Code, Anthropic's official CLI for Claude, running within the Claude Agent SDK."
   },
+  claudeOtherSystemPrompt5: {
+    category: 'system',
+    title: 'Claude CLI Billing Header',
+    text: 'x-anthropic-billing-header: cc_version=2.1.15.c5a; cc_entrypoint=cli'
+  },
   claudeOtherSystemPromptCompact: {
     category: 'system',
     title: 'Claude Code Compact System Prompt',
     text: 'You are a helpful AI assistant tasked with summarizing conversations.'
   },
+  exploreAgentSystemPrompt: {
+    category: 'system',
+    title: 'Claude Code Explore Agent System Prompt',
+    text: "You are a file search specialist for Claude Code, Anthropic's official CLI for Claude."
+  },
   outputStyleInsightsPrompt: {
     category: 'output_style',
     title: 'Output Style Insights Addendum',

src/utils/errorSanitizer.js

@@ -55,16 +55,69 @@ function sanitizeUpstreamError(errorData) {
     return errorData
   }
 
-  // 深拷贝避免修改原始对象
-  const sanitized = JSON.parse(JSON.stringify(errorData))
+  // AxiosError / Error：返回摘要，避免泄露请求体/headers/token 等敏感信息
+  const looksLikeAxiosError =
+    errorData.isAxiosError ||
+    (errorData.name === 'AxiosError' && (errorData.config || errorData.response))
+  const looksLikeError = errorData instanceof Error || typeof errorData.message === 'string'
+
+  if (looksLikeAxiosError || looksLikeError) {
+    const statusCode = errorData.response?.status
+    const upstreamBody = errorData.response?.data
+    const upstreamMessage = sanitizeErrorMessage(extractErrorMessage(upstreamBody) || '')
+
+    return {
+      name: errorData.name || 'Error',
+      code: errorData.code,
+      statusCode,
+      message: sanitizeErrorMessage(errorData.message || ''),
+      upstreamMessage: upstreamMessage || undefined,
+      upstreamType: upstreamBody?.error?.type || upstreamBody?.error?.status || undefined
+    }
+  }
 
   // 递归清理嵌套的错误对象
+  const visited = new WeakSet()
+
+  const shouldRedactKey = (key) => {
+    if (!key) {
+      return false
+    }
+    const lowerKey = String(key).toLowerCase()
+    return (
+      lowerKey === 'authorization' ||
+      lowerKey === 'cookie' ||
+      lowerKey.includes('api_key') ||
+      lowerKey.includes('apikey') ||
+      lowerKey.includes('access_token') ||
+      lowerKey.includes('refresh_token') ||
+      lowerKey.endsWith('token') ||
+      lowerKey.includes('secret') ||
+      lowerKey.includes('password')
+    )
+  }
+
   const sanitizeObject = (obj) => {
     if (!obj || typeof obj !== 'object') {
       return obj
     }
 
+    if (visited.has(obj)) {
+      return '[Circular]'
+    }
+    visited.add(obj)
+
+    // 主动剔除常见“超大且敏感”的字段
+    if (obj.config || obj.request || obj.response) {
+      return '[Redacted]'
+    }
+
     for (const key in obj) {
+      if (shouldRedactKey(key)) {
+        obj[key] = '[REDACTED]'
+        continue
+      }
+
       // 清理所有字符串字段，不仅仅是 message
       if (typeof obj[key] === 'string') {
         obj[key] = sanitizeErrorMessage(obj[key])
@@ -76,7 +129,9 @@ function sanitizeUpstreamError(errorData) {
     return obj
   }
 
-  return sanitizeObject(sanitized)
+  // 尽量不修改原对象：浅拷贝后递归清理
+  const clone = Array.isArray(errorData) ? [...errorData] : { ...errorData }
+  return sanitizeObject(clone)
 }
 
 /**

src/utils/featureFlags.js

@@ -0,0 +1,46 @@
+let config = {}
+try {
+  // config/config.js 可能在某些环境不存在（例如仅拷贝了 config.example.js）
+  // 为保证可运行，这里做容错处理
+  // eslint-disable-next-line global-require
+  config = require('../../config/config')
+} catch (error) {
+  config = {}
+}
+
+const parseBooleanEnv = (value) => {
+  if (typeof value === 'boolean') {
+    return value
+  }
+  if (typeof value !== 'string') {
+    return false
+  }
+  const normalized = value.trim().toLowerCase()
+  return normalized === 'true' || normalized === '1' || normalized === 'yes' || normalized === 'on'
+}
+
+/**
+ * 是否允许执行"余额脚本"（安全开关）
+ * ⚠️ 安全警告：vm模块非安全沙箱，默认禁用。如需启用请显式设置 BALANCE_SCRIPT_ENABLED=true
+ * 仅在完全信任管理员且了解RCE风险时才启用此功能
+ */
+const isBalanceScriptEnabled = () => {
+  if (
+    process.env.BALANCE_SCRIPT_ENABLED !== undefined &&
+    process.env.BALANCE_SCRIPT_ENABLED !== ''
+  ) {
+    return parseBooleanEnv(process.env.BALANCE_SCRIPT_ENABLED)
+  }
+
+  const fromConfig =
+    config?.accountBalance?.enableBalanceScript ??
+    config?.features?.balanceScriptEnabled ??
+    config?.security?.enableBalanceScript
+
+  // 默认禁用，需显式启用
+  return typeof fromConfig === 'boolean' ? fromConfig : false
+}
+
+module.exports = {
+  isBalanceScriptEnabled
+}

src/utils/geminiSchemaCleaner.js

@@ -0,0 +1,265 @@
+function appendHint(description, hint) {
+  if (!hint) {
+    return description || ''
+  }
+  if (!description) {
+    return hint
+  }
+  return `${description} (${hint})`
+}
+
+function getRefHint(refValue) {
+  const ref = String(refValue || '')
+  if (!ref) {
+    return ''
+  }
+  const idx = ref.lastIndexOf('/')
+  const name = idx >= 0 ? ref.slice(idx + 1) : ref
+  return name ? `See: ${name}` : ''
+}
+
+function normalizeType(typeValue) {
+  if (typeof typeValue === 'string' && typeValue) {
+    return { type: typeValue, hint: '' }
+  }
+  if (!Array.isArray(typeValue) || typeValue.length === 0) {
+    return { type: '', hint: '' }
+  }
+  const raw = typeValue.map((t) => (t === null || t === undefined ? '' : String(t))).filter(Boolean)
+  const hasNull = raw.includes('null')
+  const nonNull = raw.filter((t) => t !== 'null')
+  const primary = nonNull[0] || 'string'
+  const hintParts = []
+  if (nonNull.length > 1) {
+    hintParts.push(`Accepts: ${nonNull.join(' | ')}`)
+  }
+  if (hasNull) {
+    hintParts.push('nullable')
+  }
+  return { type: primary, hint: hintParts.join('; ') }
+}
+
+const CONSTRAINT_KEYS = [
+  'minLength',
+  'maxLength',
+  'exclusiveMinimum',
+  'exclusiveMaximum',
+  'pattern',
+  'minItems',
+  'maxItems'
+]
+
+function scoreSchema(schema) {
+  if (!schema || typeof schema !== 'object') {
+    return { score: 0, type: '' }
+  }
+  const t = typeof schema.type === 'string' ? schema.type : ''
+  if (t === 'object' || (schema.properties && typeof schema.properties === 'object')) {
+    return { score: 3, type: t || 'object' }
+  }
+  if (t === 'array' || schema.items) {
+    return { score: 2, type: t || 'array' }
+  }
+  if (t && t !== 'null') {
+    return { score: 1, type: t }
+  }
+  return { score: 0, type: t || 'null' }
+}
+
+function pickBestFromAlternatives(alternatives) {
+  let bestIndex = 0
+  let bestScore = -1
+  const types = []
+  for (let i = 0; i < alternatives.length; i += 1) {
+    const alt = alternatives[i]
+    const { score, type } = scoreSchema(alt)
+    if (type) {
+      types.push(type)
+    }
+    if (score > bestScore) {
+      bestScore = score
+      bestIndex = i
+    }
+  }
+  return { best: alternatives[bestIndex], types: Array.from(new Set(types)).filter(Boolean) }
+}
+
+function cleanJsonSchemaForGemini(schema) {
+  if (schema === null || schema === undefined) {
+    return { type: 'object', properties: {} }
+  }
+  if (typeof schema !== 'object') {
+    return { type: 'object', properties: {} }
+  }
+  if (Array.isArray(schema)) {
+    return { type: 'object', properties: {} }
+  }
+
+  // $ref：Gemini/Antigravity 不支持，转换为 hint
+  if (typeof schema.$ref === 'string' && schema.$ref) {
+    return {
+      type: 'object',
+      description: appendHint(schema.description || '', getRefHint(schema.$ref)),
+      properties: {}
+    }
+  }
+
+  // anyOf / oneOf：选择最可能的 schema，保留类型提示
+  const anyOf = Array.isArray(schema.anyOf) ? schema.anyOf : null
+  const oneOf = Array.isArray(schema.oneOf) ? schema.oneOf : null
+  const alts = anyOf && anyOf.length ? anyOf : oneOf && oneOf.length ? oneOf : null
+  if (alts) {
+    const { best, types } = pickBestFromAlternatives(alts)
+    const cleaned = cleanJsonSchemaForGemini(best)
+    const mergedDescription = appendHint(cleaned.description || '', schema.description || '')
+    const typeHint = types.length > 1 ? `Accepts: ${types.join(' || ')}` : ''
+    return {
+      ...cleaned,
+      description: appendHint(mergedDescription, typeHint)
+    }
+  }
+
+  // allOf：合并 properties/required
+  if (Array.isArray(schema.allOf) && schema.allOf.length) {
+    const merged = {}
+    let mergedDesc = schema.description || ''
+    const mergedReq = new Set()
+    const mergedProps = {}
+    for (const item of schema.allOf) {
+      const cleaned = cleanJsonSchemaForGemini(item)
+      if (cleaned.description) {
+        mergedDesc = appendHint(mergedDesc, cleaned.description)
+      }
+      if (Array.isArray(cleaned.required)) {
+        for (const r of cleaned.required) {
+          if (typeof r === 'string' && r) {
+            mergedReq.add(r)
+          }
+        }
+      }
+      if (cleaned.properties && typeof cleaned.properties === 'object') {
+        Object.assign(mergedProps, cleaned.properties)
+      }
+      if (cleaned.type && !merged.type) {
+        merged.type = cleaned.type
+      }
+      if (cleaned.items && !merged.items) {
+        merged.items = cleaned.items
+      }
+      if (Array.isArray(cleaned.enum) && !merged.enum) {
+        merged.enum = cleaned.enum
+      }
+    }
+    if (Object.keys(mergedProps).length) {
+      merged.type = merged.type || 'object'
+      merged.properties = mergedProps
+      const req = Array.from(mergedReq).filter((r) => mergedProps[r])
+      if (req.length) {
+        merged.required = req
+      }
+    }
+    if (mergedDesc) {
+      merged.description = mergedDesc
+    }
+    return cleanJsonSchemaForGemini(merged)
+  }
+
+  const result = {}
+  const constraintHints = []
+
+  // description
+  if (typeof schema.description === 'string') {
+    result.description = schema.description
+  }
+
+  for (const key of CONSTRAINT_KEYS) {
+    const value = schema[key]
+    if (value === undefined || value === null || typeof value === 'object') {
+      continue
+    }
+    constraintHints.push(`${key}: ${value}`)
+  }
+
+  // const -> enum
+  if (schema.const !== undefined && !Array.isArray(schema.enum)) {
+    result.enum = [schema.const]
+  }
+
+  // enum
+  if (Array.isArray(schema.enum)) {
+    const en = schema.enum.filter(
+      (v) => typeof v === 'string' || typeof v === 'number' || typeof v === 'boolean'
+    )
+    if (en.length) {
+      result.enum = en
+    }
+  }
+
+  // type（flatten 数组 type）
+  const { type: normalizedType, hint: typeHint } = normalizeType(schema.type)
+  if (normalizedType) {
+    result.type = normalizedType
+  }
+  if (typeHint) {
+    result.description = appendHint(result.description || '', typeHint)
+  }
+
+  if (result.enum && result.enum.length > 1 && result.enum.length <= 10) {
+    const list = result.enum.map((item) => String(item)).join(', ')
+    result.description = appendHint(result.description || '', `Allowed: ${list}`)
+  }
+
+  if (constraintHints.length) {
+    result.description = appendHint(result.description || '', constraintHints.join(', '))
+  }
+
+  // additionalProperties：Gemini/Antigravity 不接受布尔值，直接删除并用 hint 记录
+  if (schema.additionalProperties === false) {
+    result.description = appendHint(result.description || '', 'No extra properties allowed')
+  }
+
+  // properties
+  if (
+    schema.properties &&
+    typeof schema.properties === 'object' &&
+    !Array.isArray(schema.properties)
+  ) {
+    const props = {}
+    for (const [name, propSchema] of Object.entries(schema.properties)) {
+      props[name] = cleanJsonSchemaForGemini(propSchema)
+    }
+    result.type = result.type || 'object'
+    result.properties = props
+  }
+
+  // items
+  if (schema.items !== undefined) {
+    result.type = result.type || 'array'
+    result.items = cleanJsonSchemaForGemini(schema.items)
+  }
+
+  // required（最后再清理无效字段）
+  if (Array.isArray(schema.required) && result.properties) {
+    const req = schema.required.filter(
+      (r) =>
+        typeof r === 'string' && r && Object.prototype.hasOwnProperty.call(result.properties, r)
+    )
+    if (req.length) {
+      result.required = req
+    }
+  }
+
+  // 只保留 Gemini 兼容字段：其他（$schema/$id/$defs/definitions/format/constraints/pattern...）一律丢弃
+
+  if (!result.type) {
+    result.type = result.properties ? 'object' : result.items ? 'array' : 'object'
+  }
+  if (result.type === 'object' && !result.properties) {
+    result.properties = {}
+  }
+  return result
+}
+
+module.exports = {
+  cleanJsonSchemaForGemini
+}

src/utils/headerFilter.js

@@ -0,0 +1,121 @@
+/**
+ * 统一的 CDN Headers 过滤列表
+ *
+ * 用于各服务在原有过滤逻辑基础上，额外移除 Cloudflare CDN 和代理相关的 headers
+ * 避免触发上游 API（如 88code）的安全检查
+ */
+
+// Cloudflare CDN headers（橙色云代理模式会添加这些）
+const cdnHeaders = [
+  'x-real-ip',
+  'x-forwarded-for',
+  'x-forwarded-proto',
+  'x-forwarded-host',
+  'x-forwarded-port',
+  'x-accel-buffering',
+  'cf-ray',
+  'cf-connecting-ip',
+  'cf-ipcountry',
+  'cf-visitor',
+  'cf-request-id',
+  'cdn-loop',
+  'true-client-ip'
+]
+
+/**
+ * 为 OpenAI/Responses API 过滤 headers
+ * 在原有 skipHeaders 基础上添加 CDN headers
+ */
+function filterForOpenAI(headers) {
+  const skipHeaders = [
+    'host',
+    'content-length',
+    'authorization',
+    'x-api-key',
+    'x-cr-api-key',
+    'connection',
+    'upgrade',
+    'sec-websocket-key',
+    'sec-websocket-version',
+    'sec-websocket-extensions',
+    ...cdnHeaders // 添加 CDN headers
+  ]
+
+  const filtered = {}
+  for (const [key, value] of Object.entries(headers)) {
+    if (!skipHeaders.includes(key.toLowerCase())) {
+      filtered[key] = value
+    }
+  }
+  return filtered
+}
+
+/**
+ * 为 Claude/Anthropic API 过滤 headers
+ * 使用白名单模式，只允许指定的 headers 通过
+ */
+function filterForClaude(headers) {
+  // 白名单模式：只允许以下 headers
+  const allowedHeaders = [
+    'accept',
+    'x-stainless-retry-count',
+    'x-stainless-timeout',
+    'x-stainless-lang',
+    'x-stainless-package-version',
+    'x-stainless-os',
+    'x-stainless-arch',
+    'x-stainless-runtime',
+    'x-stainless-runtime-version',
+    'x-stainless-helper-method',
+    'anthropic-dangerous-direct-browser-access',
+    'anthropic-version',
+    'x-app',
+    'anthropic-beta',
+    'accept-language',
+    'sec-fetch-mode',
+    'accept-encoding',
+    'user-agent',
+    'content-type',
+    'connection'
+  ]
+
+  const filtered = {}
+  Object.keys(headers || {}).forEach((key) => {
+    const lowerKey = key.toLowerCase()
+    if (allowedHeaders.includes(lowerKey)) {
+      filtered[key] = headers[key]
+    }
+  })
+
+  return filtered
+}
+
+/**
+ * 为 Gemini API 过滤 headers（如果需要转发客户端 headers 时使用）
+ * 目前 Gemini 服务不转发客户端 headers，仅提供此方法备用
+ */
+function filterForGemini(headers) {
+  const skipHeaders = [
+    'host',
+    'content-length',
+    'authorization',
+    'x-api-key',
+    'connection',
+    ...cdnHeaders // 添加 CDN headers
+  ]
+
+  const filtered = {}
+  for (const [key, value] of Object.entries(headers)) {
+    if (!skipHeaders.includes(key.toLowerCase())) {
+      filtered[key] = value
+    }
+  }
+  return filtered
+}
+
+module.exports = {
+  cdnHeaders,
+  filterForOpenAI,
+  filterForClaude,
+  filterForGemini
+}

src/utils/logger.js

@@ -137,6 +137,7 @@ const createLogFormat = (colorize = false) => {
 
 const logFormat = createLogFormat(false)
 const consoleFormat = createLogFormat(true)
+const isTestEnv = process.env.NODE_ENV === 'test' || process.env.JEST_WORKER_ID
 
 // 📁 确保日志目录存在并设置权限
 if (!fs.existsSync(config.logging.dirname)) {
@@ -159,18 +160,20 @@ const createRotateTransport = (filename, level = null) => {
     transport.level = level
   }
 
-  // 监听轮转事件
-  transport.on('rotate', (oldFilename, newFilename) => {
-    console.log(`📦 Log rotated: ${oldFilename} -> ${newFilename}`)
-  })
+  // 监听轮转事件（测试环境关闭以避免 Jest 退出后输出）
+  if (!isTestEnv) {
+    transport.on('rotate', (oldFilename, newFilename) => {
+      console.log(`📦 Log rotated: ${oldFilename} -> ${newFilename}`)
+    })
 
-  transport.on('new', (newFilename) => {
-    console.log(`📄 New log file created: ${newFilename}`)
-  })
+    transport.on('new', (newFilename) => {
+      console.log(`📄 New log file created: ${newFilename}`)
+    })
 
-  transport.on('archive', (zipFilename) => {
-    console.log(`🗜️ Log archived: ${zipFilename}`)
-  })
+    transport.on('archive', (zipFilename) => {
+      console.log(`🗜️ Log archived: ${zipFilename}`)
+    })
+  }
 
   return transport
 }

src/utils/modelHelper.js

@@ -5,6 +5,10 @@
  * Supports parsing model strings like "ccr,model_name" to extract vendor type and base model.
  */
 
+// 仅保留原仓库既有的模型前缀：CCR 路由
+// Gemini/Antigravity 采用“路径分流”，避免在 model 字段里混入 vendor 前缀造成混乱
+const SUPPORTED_VENDOR_PREFIXES = ['ccr']
+
 /**
  * Parse vendor-prefixed model string
  * @param {string} modelStr - Model string, potentially with vendor prefix (e.g., "ccr,gemini-2.5-pro")
@@ -19,16 +23,21 @@ function parseVendorPrefixedModel(modelStr) {
   const trimmed = modelStr.trim()
   const lowerTrimmed = trimmed.toLowerCase()
 
-  // Check for ccr prefix (case insensitive)
-  if (lowerTrimmed.startsWith('ccr,')) {
+  for (const vendorPrefix of SUPPORTED_VENDOR_PREFIXES) {
+    if (!lowerTrimmed.startsWith(`${vendorPrefix},`)) {
+      continue
+    }
+
     const parts = trimmed.split(',')
-    if (parts.length >= 2) {
-      // Extract base model (everything after the first comma, rejoined in case model name contains commas)
-      const baseModel = parts.slice(1).join(',').trim()
-      return {
-        vendor: 'ccr',
-        baseModel
-      }
+    if (parts.length < 2) {
+      break
+    }
+
+    // Extract base model (everything after the first comma, rejoined in case model name contains commas)
+    const baseModel = parts.slice(1).join(',').trim()
+    return {
+      vendor: vendorPrefix,
+      baseModel
     }
   }
 
@@ -70,9 +79,119 @@ function getVendorType(modelStr) {
   return vendor
 }
 
+/**
+ * Check if the model is Opus 4.5 or newer.
+ *
+ * VERSION LOGIC (as of 2025-12-05):
+ * - Opus 4.5+ (including 5.0, 6.0, etc.) → returns true (Pro account eligible)
+ * - Opus 4.4 and below (including 3.x, 4.0, 4.1) → returns false (Max account only)
+ *
+ * Supported naming formats:
+ *   - New format: claude-opus-{major}[-{minor}][-date], e.g., claude-opus-4-5-20251101
+ *   - New format: claude-opus-{major}.{minor}, e.g., claude-opus-4.5
+ *   - Old format: claude-{version}-opus[-date], e.g., claude-3-opus-20240229
+ *   - Special: opus-latest, claude-opus-latest → always returns true
+ *
+ * @param {string} modelName - Model name
+ * @returns {boolean} - Whether the model is Opus 4.5 or newer
+ */
+function isOpus45OrNewer(modelName) {
+  if (!modelName) {
+    return false
+  }
+
+  const lowerModel = modelName.toLowerCase()
+  if (!lowerModel.includes('opus')) {
+    return false
+  }
+
+  // Handle 'latest' special case
+  if (lowerModel.includes('opus-latest') || lowerModel.includes('opus_latest')) {
+    return true
+  }
+
+  // Old format: claude-{version}-opus (version before opus)
+  // e.g., claude-3-opus-20240229, claude-3.5-opus
+  const oldFormatMatch = lowerModel.match(/claude[- ](\d+)(?:[.-](\d+))?[- ]opus/)
+  if (oldFormatMatch) {
+    const majorVersion = parseInt(oldFormatMatch[1], 10)
+    const minorVersion = oldFormatMatch[2] ? parseInt(oldFormatMatch[2], 10) : 0
+
+    // Old format version refers to Claude major version
+    // majorVersion > 4: 5.x, 6.x, ... → true
+    // majorVersion === 4 && minorVersion >= 5: 4.5, 4.6, ... → true
+    // Others (3.x, 4.0-4.4): → false
+    if (majorVersion > 4) {
+      return true
+    }
+    if (majorVersion === 4 && minorVersion >= 5) {
+      return true
+    }
+    return false
+  }
+
+  // New format 1: opus-{major}.{minor} (dot-separated)
+  // e.g., claude-opus-4.5, opus-4.5
+  const dotFormatMatch = lowerModel.match(/opus[- ]?(\d+)\.(\d+)/)
+  if (dotFormatMatch) {
+    const majorVersion = parseInt(dotFormatMatch[1], 10)
+    const minorVersion = parseInt(dotFormatMatch[2], 10)
+
+    // Same version logic as old format
+    // opus-5.0, opus-6.0 → true
+    // opus-4.5, opus-4.6 → true
+    // opus-4.0, opus-4.4 → false
+    if (majorVersion > 4) {
+      return true
+    }
+    if (majorVersion === 4 && minorVersion >= 5) {
+      return true
+    }
+    return false
+  }
+
+  // New format 2: opus-{major}[-{minor}][-date] (hyphen-separated)
+  // e.g., claude-opus-4-5-20251101, claude-opus-4-20250514, claude-opus-4-1-20250805
+  // If opus-{major} is followed by 8-digit date, there's no minor version
+
+  // Extract content after 'opus'
+  const opusIndex = lowerModel.indexOf('opus')
+  const afterOpus = lowerModel.substring(opusIndex + 4)
+
+  // Match: -{major}-{minor}-{date} or -{major}-{date} or -{major}
+  // IMPORTANT: Minor version regex is (\d{1,2}) not (\d+)
+  // This prevents matching 8-digit dates as minor version
+  // Example: opus-4-20250514 → major=4, minor=undefined (not 20250514)
+  // Example: opus-4-5-20251101 → major=4, minor=5
+  // Future-proof: Supports up to 2-digit minor versions (0-99)
+  const versionMatch = afterOpus.match(/^[- ](\d+)(?:[- ](\d{1,2})(?=[- ]\d{8}|$))?/)
+
+  if (versionMatch) {
+    const majorVersion = parseInt(versionMatch[1], 10)
+    const minorVersion = versionMatch[2] ? parseInt(versionMatch[2], 10) : 0
+
+    // Same version logic: >= 4.5 returns true
+    // opus-5-0-date, opus-6-date → true
+    // opus-4-5-date, opus-4-10-date → true (supports 2-digit minor)
+    // opus-4-date (no minor, treated as 4.0) → false
+    // opus-4-1-date, opus-4-4-date → false
+    if (majorVersion > 4) {
+      return true
+    }
+    if (majorVersion === 4 && minorVersion >= 5) {
+      return true
+    }
+    return false
+  }
+
+  // Other cases containing 'opus' but cannot parse version, assume legacy
+  return false
+}
+
 module.exports = {
   parseVendorPrefixedModel,
   hasVendorPrefix,
   getEffectiveModel,
-  getVendorType
+  getVendorType,
+  isOpus45OrNewer
 }

src/utils/oauthHelper.js

@@ -13,11 +13,18 @@ const OAUTH_CONFIG = {
   AUTHORIZE_URL: 'https://claude.ai/oauth/authorize',
   TOKEN_URL: 'https://console.anthropic.com/v1/oauth/token',
   CLIENT_ID: '9d1c250a-e61b-44d9-88ed-5944d1962f5e',
-  REDIRECT_URI: 'https://console.anthropic.com/oauth/code/callback',
-  SCOPES: 'org:create_api_key user:profile user:inference',
+  REDIRECT_URI: 'https://platform.claude.com/oauth/code/callback',
+  SCOPES: 'org:create_api_key user:profile user:inference user:sessions:claude_code',
   SCOPES_SETUP: 'user:inference' // Setup Token 只需要推理权限
 }
 
+// Cookie自动授权配置常量
+const COOKIE_OAUTH_CONFIG = {
+  CLAUDE_AI_URL: 'https://claude.ai',
+  ORGANIZATIONS_URL: 'https://claude.ai/api/organizations',
+  AUTHORIZE_URL_TEMPLATE: 'https://claude.ai/v1/oauth/{organization_uuid}/authorize'
+}
+
 /**
  * 生成随机的 state 参数
  * @returns {string} 随机生成的 state (base64url编码)
@@ -28,6 +35,7 @@ function generateState() {
 
 /**
  * 生成随机的 code verifier（PKCE）
+ * 符合 RFC 7636 标准：32字节随机数 → base64url编码 → 43字符
  * @returns {string} base64url 编码的随机字符串
  */
 function generateCodeVerifier() {
@@ -570,8 +578,299 @@ function extractExtInfo(data) {
   return Object.keys(ext).length > 0 ? ext : null
 }
 
+// =============================================================================
+// Cookie自动授权相关方法 (基于Clove项目实现)
+// =============================================================================
+
+/**
+ * 构建带Cookie的请求头
+ * @param {string} sessionKey - sessionKey值
+ * @returns {object} 请求头对象
+ */
+function buildCookieHeaders(sessionKey) {
+  return {
+    Accept: 'application/json',
+    'Accept-Language': 'en-US,en;q=0.9',
+    'Cache-Control': 'no-cache',
+    Cookie: `sessionKey=${sessionKey}`,
+    Origin: COOKIE_OAUTH_CONFIG.CLAUDE_AI_URL,
+    Referer: `${COOKIE_OAUTH_CONFIG.CLAUDE_AI_URL}/new`,
+    'User-Agent':
+      'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36'
+  }
+}
+
+/**
+ * 使用Cookie获取组织UUID和能力列表
+ * @param {string} sessionKey - sessionKey值
+ * @param {object|null} proxyConfig - 代理配置（可选）
+ * @returns {Promise<{organizationUuid: string, capabilities: string[]}>}
+ */
+async function getOrganizationInfo(sessionKey, proxyConfig = null) {
+  const headers = buildCookieHeaders(sessionKey)
+  const agent = createProxyAgent(proxyConfig)
+
+  try {
+    if (agent) {
+      logger.info(`🌐 Using proxy for organization info: ${ProxyHelper.maskProxyInfo(proxyConfig)}`)
+    }
+
+    logger.debug('🔄 Fetching organization info with Cookie', {
+      url: COOKIE_OAUTH_CONFIG.ORGANIZATIONS_URL,
+      hasProxy: !!proxyConfig
+    })
+
+    const axiosConfig = {
+      headers,
+      timeout: 30000,
+      maxRedirects: 0 // 禁止自动重定向，以便检测Cloudflare拦截(302)
+    }
+
+    if (agent) {
+      axiosConfig.httpAgent = agent
+      axiosConfig.httpsAgent = agent
+      axiosConfig.proxy = false
+    }
+
+    const response = await axios.get(COOKIE_OAUTH_CONFIG.ORGANIZATIONS_URL, axiosConfig)
+
+    if (!response.data || !Array.isArray(response.data)) {
+      throw new Error('获取组织信息失败：响应格式无效')
+    }
+
+    // 找到具有chat能力且能力最多的组织
+    let bestOrg = null
+    let maxCapabilities = []
+
+    for (const org of response.data) {
+      const capabilities = org.capabilities || []
+
+      // 必须有chat能力
+      if (!capabilities.includes('chat')) {
+        continue
+      }
+
+      // 选择能力最多的组织
+      if (capabilities.length > maxCapabilities.length) {
+        bestOrg = org
+        maxCapabilities = capabilities
+      }
+    }
+
+    if (!bestOrg || !bestOrg.uuid) {
+      throw new Error('未找到具有chat能力的组织')
+    }
+
+    logger.success('✅ Found organization', {
+      uuid: bestOrg.uuid,
+      capabilities: maxCapabilities
+    })
+
+    return {
+      organizationUuid: bestOrg.uuid,
+      capabilities: maxCapabilities
+    }
+  } catch (error) {
+    if (error.response) {
+      const { status } = error.response
+
+      if (status === 403 || status === 401) {
+        throw new Error('Cookie授权失败：无效的sessionKey或已过期')
+      }
+
+      if (status === 302) {
+        throw new Error('请求被Cloudflare拦截，请稍后重试')
+      }
+
+      throw new Error(`获取组织信息失败：HTTP ${status}`)
+    } else if (error.request) {
+      throw new Error('获取组织信息失败：网络错误或超时')
+    }
+
+    throw error
+  }
+}
+
+/**
+ * 使用Cookie自动获取授权code
+ * @param {string} sessionKey - sessionKey值
+ * @param {string} organizationUuid - 组织UUID
+ * @param {string} scope - 授权scope
+ * @param {object|null} proxyConfig - 代理配置（可选）
+ * @returns {Promise<{authorizationCode: string, codeVerifier: string, state: string}>}
+ */
+async function authorizeWithCookie(sessionKey, organizationUuid, scope, proxyConfig = null) {
+  // 生成PKCE参数
+  const codeVerifier = generateCodeVerifier()
+  const codeChallenge = generateCodeChallenge(codeVerifier)
+  const state = generateState()
+
+  // 构建授权URL
+  const authorizeUrl = COOKIE_OAUTH_CONFIG.AUTHORIZE_URL_TEMPLATE.replace(
+    '{organization_uuid}',
+    organizationUuid
+  )
+
+  // 构建请求payload
+  const payload = {
+    response_type: 'code',
+    client_id: OAUTH_CONFIG.CLIENT_ID,
+    organization_uuid: organizationUuid,
+    redirect_uri: OAUTH_CONFIG.REDIRECT_URI,
+    scope,
+    state,
+    code_challenge: codeChallenge,
+    code_challenge_method: 'S256'
+  }
+
+  const headers = {
+    ...buildCookieHeaders(sessionKey),
+    'Content-Type': 'application/json'
+  }
+
+  const agent = createProxyAgent(proxyConfig)
+
+  try {
+    if (agent) {
+      logger.info(
+        `🌐 Using proxy for Cookie authorization: ${ProxyHelper.maskProxyInfo(proxyConfig)}`
+      )
+    }
+
+    logger.debug('🔄 Requesting authorization with Cookie', {
+      url: authorizeUrl,
+      scope,
+      hasProxy: !!proxyConfig
+    })
+
+    const axiosConfig = {
+      headers,
+      timeout: 30000,
+      maxRedirects: 0 // 禁止自动重定向，以便检测Cloudflare拦截(302)
+    }
+
+    if (agent) {
+      axiosConfig.httpAgent = agent
+      axiosConfig.httpsAgent = agent
+      axiosConfig.proxy = false
+    }
+
+    const response = await axios.post(authorizeUrl, payload, axiosConfig)
+
+    // 从响应中获取redirect_uri
+    const redirectUri = response.data?.redirect_uri
+
+    if (!redirectUri) {
+      throw new Error('授权响应中未找到redirect_uri')
+    }
+
+    logger.debug('📎 Got redirect URI', { redirectUri: `${redirectUri.substring(0, 80)}...` })
+
+    // 解析redirect_uri获取authorization code
+    const url = new URL(redirectUri)
+    const authorizationCode = url.searchParams.get('code')
+    const responseState = url.searchParams.get('state')
+
+    if (!authorizationCode) {
+      throw new Error('redirect_uri中未找到授权码')
+    }
+
+    // 构建完整的授权码（包含state，如果有的话）
+    const fullCode = responseState ? `${authorizationCode}#${responseState}` : authorizationCode
+
+    logger.success('✅ Got authorization code via Cookie', {
+      codeLength: authorizationCode.length,
+      codePrefix: `${authorizationCode.substring(0, 10)}...`
+    })
+
+    return {
+      authorizationCode: fullCode,
+      codeVerifier,
+      state
+    }
+  } catch (error) {
+    if (error.response) {
+      const { status } = error.response
+
+      if (status === 403 || status === 401) {
+        throw new Error('Cookie授权失败：无效的sessionKey或已过期')
+      }
+
+      if (status === 302) {
+        throw new Error('请求被Cloudflare拦截，请稍后重试')
+      }
+
+      const errorData = error.response.data
+      let errorMessage = `HTTP ${status}`
+
+      if (errorData) {
+        if (typeof errorData === 'string') {
+          errorMessage += `: ${errorData}`
+        } else if (errorData.error) {
+          errorMessage += `: ${errorData.error}`
+        }
+      }
+
+      throw new Error(`授权请求失败：${errorMessage}`)
+    } else if (error.request) {
+      throw new Error('授权请求失败：网络错误或超时')
+    }
+
+    throw error
+  }
+}
+
+/**
+ * 完整的Cookie自动授权流程
+ * @param {string} sessionKey - sessionKey值
+ * @param {object|null} proxyConfig - 代理配置（可选）
+ * @param {boolean} isSetupToken - 是否为Setup Token模式
+ * @returns {Promise<{claudeAiOauth: object, organizationUuid: string, capabilities: string[]}>}
+ */
+async function oauthWithCookie(sessionKey, proxyConfig = null, isSetupToken = false) {
+  logger.info('🍪 Starting Cookie-based OAuth flow', {
+    isSetupToken,
+    hasProxy: !!proxyConfig
+  })
+
+  // 步骤1：获取组织信息
+  logger.debug('Step 1/3: Fetching organization info...')
+  const { organizationUuid, capabilities } = await getOrganizationInfo(sessionKey, proxyConfig)
+
+  // 步骤2：确定scope并获取授权code
+  const scope = isSetupToken ? OAUTH_CONFIG.SCOPES_SETUP : 'user:profile user:inference'
+
+  logger.debug('Step 2/3: Getting authorization code...', { scope })
+  const { authorizationCode, codeVerifier, state } = await authorizeWithCookie(
+    sessionKey,
+    organizationUuid,
+    scope,
+    proxyConfig
+  )
+
+  // 步骤3：交换token
+  logger.debug('Step 3/3: Exchanging token...')
+  const tokenData = isSetupToken
+    ? await exchangeSetupTokenCode(authorizationCode, codeVerifier, state, proxyConfig)
+    : await exchangeCodeForTokens(authorizationCode, codeVerifier, state, proxyConfig)
+
+  logger.success('✅ Cookie-based OAuth flow completed', {
+    isSetupToken,
+    organizationUuid,
+    hasAccessToken: !!tokenData.accessToken,
+    hasRefreshToken: !!tokenData.refreshToken
+  })
+
+  return {
+    claudeAiOauth: tokenData,
+    organizationUuid,
+    capabilities
+  }
+}
+
 module.exports = {
   OAUTH_CONFIG,
+  COOKIE_OAUTH_CONFIG,
   generateOAuthParams,
   generateSetupTokenParams,
   exchangeCodeForTokens,
@@ -584,5 +883,10 @@ module.exports = {
   generateCodeChallenge,
   generateAuthUrl,
   generateSetupTokenAuthUrl,
-  createProxyAgent
+  createProxyAgent,
+  // Cookie自动授权相关方法
+  buildCookieHeaders,
+  getOrganizationInfo,
+  authorizeWithCookie,
+  oauthWithCookie
 }

src/utils/projectPaths.js

@@ -0,0 +1,10 @@
+const path = require('path')
+
+// 该文件位于 src/utils 下，向上两级即项目根目录。
+function getProjectRoot() {
+  return path.resolve(__dirname, '..', '..')
+}
+
+module.exports = {
+  getProjectRoot
+}

src/utils/safeRotatingAppend.js

@@ -0,0 +1,88 @@
+/**
+ * ============================================================================
+ * 安全 JSONL 追加工具（带文件大小限制与自动轮转）
+ * ============================================================================
+ *
+ * 用于所有调试 Dump 模块，避免日志文件无限增长导致 I/O 拥塞。
+ *
+ * 策略：
+ * - 每次写入前检查目标文件大小
+ * - 超过阈值时，将现有文件重命名为 .bak（覆盖旧 .bak）
+ * - 然后写入新文件
+ */
+
+const fs = require('fs/promises')
+const logger = require('./logger')
+
+// 默认文件大小上限：10MB
+const DEFAULT_MAX_FILE_SIZE_BYTES = 10 * 1024 * 1024
+const MAX_FILE_SIZE_ENV = 'DUMP_MAX_FILE_SIZE_BYTES'
+
+/**
+ * 获取文件大小上限（可通过环境变量覆盖）
+ */
+function getMaxFileSize() {
+  const raw = process.env[MAX_FILE_SIZE_ENV]
+  if (raw) {
+    const parsed = Number.parseInt(raw, 10)
+    if (Number.isFinite(parsed) && parsed > 0) {
+      return parsed
+    }
+  }
+  return DEFAULT_MAX_FILE_SIZE_BYTES
+}
+
+/**
+ * 获取文件大小，文件不存在时返回 0
+ */
+async function getFileSize(filepath) {
+  try {
+    const stat = await fs.stat(filepath)
+    return stat.size
+  } catch (e) {
+    // 文件不存在或无法读取
+    return 0
+  }
+}
+
+/**
+ * 安全追加写入 JSONL 文件，支持自动轮转
+ *
+ * @param {string} filepath - 目标文件绝对路径
+ * @param {string} line - 要写入的单行（应以 \n 结尾）
+ * @param {Object} options - 可选配置
+ * @param {number} options.maxFileSize - 文件大小上限（字节），默认从环境变量或 10MB
+ */
+async function safeRotatingAppend(filepath, line, options = {}) {
+  const maxFileSize = options.maxFileSize || getMaxFileSize()
+
+  const currentSize = await getFileSize(filepath)
+
+  // 如果当前文件已达到或超过阈值，轮转
+  if (currentSize >= maxFileSize) {
+    const backupPath = `${filepath}.bak`
+    try {
+      // 先删除旧备份（如果存在）
+      await fs.unlink(backupPath).catch(() => {})
+      // 重命名当前文件为备份
+      await fs.rename(filepath, backupPath)
+    } catch (renameErr) {
+      // 轮转失败时记录警告日志，继续写入原文件
+      logger.warn('⚠️ Log rotation failed, continuing to write to original file', {
+        filepath,
+        backupPath,
+        error: renameErr?.message || String(renameErr)
+      })
+    }
+  }
+
+  // 追加写入
+  await fs.appendFile(filepath, line, { encoding: 'utf8' })
+}
+
+module.exports = {
+  safeRotatingAppend,
+  getMaxFileSize,
+  MAX_FILE_SIZE_ENV,
+  DEFAULT_MAX_FILE_SIZE_BYTES
+}

src/utils/signatureCache.js

@@ -0,0 +1,183 @@
+/**
+ * Signature Cache - 签名缓存模块
+ *
+ * 用于缓存 Antigravity thinking block 的 thoughtSignature。
+ * Claude Code 客户端可能剥离非标准字段，导致多轮对话时签名丢失。
+ * 此模块按 sessionId + thinkingText 存储签名，便于后续请求恢复。
+ *
+ * 参考实现：
+ * - CLIProxyAPI: internal/cache/signature_cache.go
+ * - antigravity-claude-proxy: src/format/signature-cache.js
+ */
+
+const crypto = require('crypto')
+const logger = require('./logger')
+
+// 配置常量
+const SIGNATURE_CACHE_TTL_MS = 60 * 60 * 1000 // 1 小时（同 CLIProxyAPI）
+const MAX_ENTRIES_PER_SESSION = 100 // 每会话最大缓存条目
+const MIN_SIGNATURE_LENGTH = 50 // 最小有效签名长度
+const TEXT_HASH_LENGTH = 16 // 文本哈希长度（SHA256 前 16 位）
+
+// 主缓存：sessionId -> Map<textHash, { signature, timestamp }>
+const signatureCache = new Map()
+
+/**
+ * 生成文本内容的稳定哈希值
+ * @param {string} text - 待哈希的文本
+ * @returns {string} 16 字符的十六进制哈希
+ */
+function hashText(text) {
+  if (!text || typeof text !== 'string') {
+    return ''
+  }
+  const hash = crypto.createHash('sha256').update(text).digest('hex')
+  return hash.slice(0, TEXT_HASH_LENGTH)
+}
+
+/**
+ * 获取或创建会话缓存
+ * @param {string} sessionId - 会话 ID
+ * @returns {Map} 会话的签名缓存 Map
+ */
+function getOrCreateSessionCache(sessionId) {
+  if (!signatureCache.has(sessionId)) {
+    signatureCache.set(sessionId, new Map())
+  }
+  return signatureCache.get(sessionId)
+}
+
+/**
+ * 检查签名是否有效
+ * @param {string} signature - 待检查的签名
+ * @returns {boolean} 签名是否有效
+ */
+function isValidSignature(signature) {
+  return typeof signature === 'string' && signature.length >= MIN_SIGNATURE_LENGTH
+}
+
+/**
+ * 缓存 thinking 签名
+ * @param {string} sessionId - 会话 ID
+ * @param {string} thinkingText - thinking 内容文本
+ * @param {string} signature - thoughtSignature
+ */
+function cacheSignature(sessionId, thinkingText, signature) {
+  if (!sessionId || !thinkingText || !signature) {
+    return
+  }
+
+  if (!isValidSignature(signature)) {
+    return
+  }
+
+  const sessionCache = getOrCreateSessionCache(sessionId)
+  const textHash = hashText(thinkingText)
+
+  if (!textHash) {
+    return
+  }
+
+  // 淘汰策略：超过限制时删除最老的 1/4 条目
+  if (sessionCache.size >= MAX_ENTRIES_PER_SESSION) {
+    const entries = Array.from(sessionCache.entries())
+    entries.sort((a, b) => a[1].timestamp - b[1].timestamp)
+    const toRemove = Math.max(1, Math.floor(entries.length / 4))
+    for (let i = 0; i < toRemove; i++) {
+      sessionCache.delete(entries[i][0])
+    }
+    logger.debug(
+      `[SignatureCache] Evicted ${toRemove} old entries for session ${sessionId.slice(0, 8)}...`
+    )
+  }
+
+  sessionCache.set(textHash, {
+    signature,
+    timestamp: Date.now()
+  })
+
+  logger.debug(
+    `[SignatureCache] Cached signature for session ${sessionId.slice(0, 8)}..., hash ${textHash}`
+  )
+}
+
+/**
+ * 获取缓存的签名
+ * @param {string} sessionId - 会话 ID
+ * @param {string} thinkingText - thinking 内容文本
+ * @returns {string|null} 缓存的签名，未找到或过期则返回 null
+ */
+function getCachedSignature(sessionId, thinkingText) {
+  if (!sessionId || !thinkingText) {
+    return null
+  }
+
+  const sessionCache = signatureCache.get(sessionId)
+  if (!sessionCache) {
+    return null
+  }
+
+  const textHash = hashText(thinkingText)
+  if (!textHash) {
+    return null
+  }
+
+  const entry = sessionCache.get(textHash)
+  if (!entry) {
+    return null
+  }
+
+  // 检查是否过期
+  if (Date.now() - entry.timestamp > SIGNATURE_CACHE_TTL_MS) {
+    sessionCache.delete(textHash)
+    logger.debug(`[SignatureCache] Entry expired for hash ${textHash}`)
+    return null
+  }
+
+  logger.debug(
+    `[SignatureCache] Cache hit for session ${sessionId.slice(0, 8)}..., hash ${textHash}`
+  )
+  return entry.signature
+}
+
+/**
+ * 清除会话缓存
+ * @param {string} sessionId - 要清除的会话 ID，为空则清除全部
+ */
+function clearSignatureCache(sessionId = null) {
+  if (sessionId) {
+    signatureCache.delete(sessionId)
+    logger.debug(`[SignatureCache] Cleared cache for session ${sessionId.slice(0, 8)}...`)
+  } else {
+    signatureCache.clear()
+    logger.debug('[SignatureCache] Cleared all caches')
+  }
+}
+
+/**
+ * 获取缓存统计信息（调试用）
+ * @returns {Object} { sessionCount, totalEntries }
+ */
+function getCacheStats() {
+  let totalEntries = 0
+  for (const sessionCache of signatureCache.values()) {
+    totalEntries += sessionCache.size
+  }
+  return {
+    sessionCount: signatureCache.size,
+    totalEntries
+  }
+}
+
+module.exports = {
+  cacheSignature,
+  getCachedSignature,
+  clearSignatureCache,
+  getCacheStats,
+  isValidSignature,
+  // 内部函数导出（用于测试或扩展）
+  hashText,
+  MIN_SIGNATURE_LENGTH,
+  MAX_ENTRIES_PER_SESSION,
+  SIGNATURE_CACHE_TTL_MS
+}

src/utils/statsHelper.js

@@ -0,0 +1,105 @@
+/**
+ * 统计计算工具函数
+ * 提供百分位数计算、等待时间统计等通用统计功能
+ */
+
+/**
+ * 计算百分位数（使用 nearest-rank 方法）
+ * @param {number[]} sortedArray - 已排序的数组（升序）
+ * @param {number} percentile - 百分位数 (0-100)
+ * @returns {number} 百分位值
+ *
+ * 边界情况说明：
+ * - percentile=0: 返回最小值 (index=0)
+ * - percentile=100: 返回最大值 (index=len-1)
+ * - percentile=50 且 len=2: 返回第一个元素（nearest-rank 向下取）
+ *
+ * 算法说明（nearest-rank 方法）：
+ * - index = ceil(percentile / 100 * len) - 1
+ * - 示例：len=100, P50 → ceil(50) - 1 = 49（第50个元素，0-indexed）
+ * - 示例：len=100, P99 → ceil(99) - 1 = 98（第99个元素）
+ */
+function getPercentile(sortedArray, percentile) {
+  const len = sortedArray.length
+  if (len === 0) {
+    return 0
+  }
+  if (len === 1) {
+    return sortedArray[0]
+  }
+
+  // 边界处理：percentile <= 0 返回最小值
+  if (percentile <= 0) {
+    return sortedArray[0]
+  }
+  // 边界处理：percentile >= 100 返回最大值
+  if (percentile >= 100) {
+    return sortedArray[len - 1]
+  }
+
+  const index = Math.ceil((percentile / 100) * len) - 1
+  return sortedArray[index]
+}
+
+/**
+ * 计算等待时间分布统计
+ * @param {number[]} waitTimes - 等待时间数组（无需预先排序）
+ * @returns {Object|null} 统计对象，空数组返回 null
+ *
+ * 返回对象包含：
+ * - sampleCount: 样本数量（始终包含，便于调用方判断可靠性）
+ * - count: 样本数量（向后兼容）
+ * - min: 最小值
+ * - max: 最大值
+ * - avg: 平均值（四舍五入）
+ * - p50: 50百分位数（中位数）
+ * - p90: 90百分位数
+ * - p99: 99百分位数
+ * - sampleSizeWarning: 样本量不足时的警告信息（样本 < 10）
+ * - p90Unreliable: P90 统计不可靠标记（样本 < 10）
+ * - p99Unreliable: P99 统计不可靠标记（样本 < 100）
+ *
+ * 可靠性标记说明（详见 design.md Decision 6）：
+ * - 样本 < 10: P90 和 P99 都不可靠
+ * - 样本 < 100: P99 不可靠（P90 需要 10 个样本，P99 需要 100 个样本）
+ * - 即使标记为不可靠，仍返回计算值供参考
+ */
+function calculateWaitTimeStats(waitTimes) {
+  if (!waitTimes || waitTimes.length === 0) {
+    return null
+  }
+
+  const sorted = [...waitTimes].sort((a, b) => a - b)
+  const sum = sorted.reduce((a, b) => a + b, 0)
+  const len = sorted.length
+
+  const stats = {
+    sampleCount: len, // 新增：始终包含样本数
+    count: len, // 向后兼容
+    min: sorted[0],
+    max: sorted[len - 1],
+    avg: Math.round(sum / len),
+    p50: getPercentile(sorted, 50),
+    p90: getPercentile(sorted, 90),
+    p99: getPercentile(sorted, 99)
+  }
+
+  // 渐进式可靠性标记（详见 design.md Decision 6）
+  // 样本 < 10: P90 不可靠（P90 至少需要 ceil(100/10) = 10 个样本）
+  if (len < 10) {
+    stats.sampleSizeWarning = 'Results may be inaccurate due to small sample size'
+    stats.p90Unreliable = true
+  }
+
+  // 样本 < 100: P99 不可靠（P99 至少需要 ceil(100/1) = 100 个样本）
+  if (len < 100) {
+    stats.p99Unreliable = true
+  }
+
+  return stats
+}
+
+module.exports = {
+  getPercentile,
+  calculateWaitTimeStats
+}

src/utils/streamHelper.js

@@ -0,0 +1,36 @@
+/**
+ * Stream Helper Utilities
+ * 流处理辅助工具函数
+ */
+
+/**
+ * 检查响应流是否仍然可写（客户端连接是否有效）
+ * @param {import('http').ServerResponse} stream - HTTP响应流
+ * @returns {boolean} 如果流可写返回true，否则返回false
+ */
+function isStreamWritable(stream) {
+  if (!stream) {
+    return false
+  }
+
+  // 检查流是否已销毁
+  if (stream.destroyed) {
+    return false
+  }
+
+  // 检查底层socket是否已销毁
+  if (stream.socket?.destroyed) {
+    return false
+  }
+
+  // 检查流是否已结束写入
+  if (stream.writableEnded) {
+    return false
+  }
+
+  return true
+}
+
+module.exports = {
+  isStreamWritable
+}

src/utils/testPayloadHelper.js

@@ -0,0 +1,242 @@
+const crypto = require('crypto')
+
+/**
+ * 生成随机十六进制字符串
+ * @param {number} bytes - 字节数
+ * @returns {string} 十六进制字符串
+ */
+function randomHex(bytes = 32) {
+  return crypto.randomBytes(bytes).toString('hex')
+}
+
+/**
+ * 生成 Claude Code 风格的会话字符串
+ * @returns {string} 会话字符串，格式: user_{64位hex}_account__session_{uuid}
+ */
+function generateSessionString() {
+  const hex64 = randomHex(32) // 32 bytes => 64 hex characters
+  const uuid = crypto.randomUUID()
+  return `user_${hex64}_account__session_${uuid}`
+}
+
+/**
+ * 生成 Claude 测试请求体
+ * @param {string} model - 模型名称
+ * @param {object} options - 可选配置
+ * @param {boolean} options.stream - 是否流式（默认false）
+ * @returns {object} 测试请求体
+ */
+function createClaudeTestPayload(model = 'claude-sonnet-4-5-20250929', options = {}) {
+  const payload = {
+    model,
+    messages: [
+      {
+        role: 'user',
+        content: [
+          {
+            type: 'text',
+            text: 'hi',
+            cache_control: {
+              type: 'ephemeral'
+            }
+          }
+        ]
+      }
+    ],
+    system: [
+      {
+        type: 'text',
+        text: "You are Claude Code, Anthropic's official CLI for Claude.",
+        cache_control: {
+          type: 'ephemeral'
+        }
+      }
+    ],
+    metadata: {
+      user_id: generateSessionString()
+    },
+    max_tokens: 21333,
+    temperature: 1
+  }
+
+  if (options.stream) {
+    payload.stream = true
+  }
+
+  return payload
+}
+
+/**
+ * 发送流式测试请求并处理SSE响应
+ * @param {object} options - 配置选项
+ * @param {string} options.apiUrl - API URL
+ * @param {string} options.authorization - Authorization header值
+ * @param {object} options.responseStream - Express响应流
+ * @param {object} [options.payload] - 请求体（默认使用createClaudeTestPayload）
+ * @param {object} [options.proxyAgent] - 代理agent
+ * @param {number} [options.timeout] - 超时时间（默认30000）
+ * @param {object} [options.extraHeaders] - 额外的请求头
+ * @returns {Promise<void>}
+ */
+async function sendStreamTestRequest(options) {
+  const axios = require('axios')
+  const logger = require('./logger')
+
+  const {
+    apiUrl,
+    authorization,
+    responseStream,
+    payload = createClaudeTestPayload('claude-sonnet-4-5-20250929', { stream: true }),
+    proxyAgent = null,
+    timeout = 30000,
+    extraHeaders = {}
+  } = options
+
+  const sendSSE = (type, data = {}) => {
+    if (!responseStream.destroyed && !responseStream.writableEnded) {
+      try {
+        responseStream.write(`data: ${JSON.stringify({ type, ...data })}\n\n`)
+      } catch {
+        // ignore
+      }
+    }
+  }
+
+  const endTest = (success, error = null) => {
+    if (!responseStream.destroyed && !responseStream.writableEnded) {
+      try {
+        responseStream.write(
+          `data: ${JSON.stringify({ type: 'test_complete', success, error: error || undefined })}\n\n`
+        )
+        responseStream.end()
+      } catch {
+        // ignore
+      }
+    }
+  }
+
+  // 设置响应头
+  if (!responseStream.headersSent) {
+    responseStream.writeHead(200, {
+      'Content-Type': 'text/event-stream',
+      'Cache-Control': 'no-cache',
+      Connection: 'keep-alive',
+      'X-Accel-Buffering': 'no'
+    })
+  }
+
+  sendSSE('test_start', { message: 'Test started' })
+
+  const requestConfig = {
+    method: 'POST',
+    url: apiUrl,
+    data: payload,
+    headers: {
+      'Content-Type': 'application/json',
+      'anthropic-version': '2023-06-01',
+      'User-Agent': 'claude-cli/2.0.52 (external, cli)',
+      authorization,
+      ...extraHeaders
+    },
+    timeout,
+    responseType: 'stream',
+    validateStatus: () => true
+  }
+
+  if (proxyAgent) {
+    requestConfig.httpAgent = proxyAgent
+    requestConfig.httpsAgent = proxyAgent
+    requestConfig.proxy = false
+  }
+
+  try {
+    const response = await axios(requestConfig)
+    logger.debug(`🌊 Test response status: ${response.status}`)
+
+    // 处理非200响应
+    if (response.status !== 200) {
+      return new Promise((resolve) => {
+        const chunks = []
+        response.data.on('data', (chunk) => chunks.push(chunk))
+        response.data.on('end', () => {
+          const errorData = Buffer.concat(chunks).toString()
+          let errorMsg = `API Error: ${response.status}`
+          try {
+            const json = JSON.parse(errorData)
+            errorMsg = json.message || json.error?.message || json.error || errorMsg
+          } catch {
+            if (errorData.length < 200) {
+              errorMsg = errorData || errorMsg
+            }
+          }
+          endTest(false, errorMsg)
+          resolve()
+        })
+        response.data.on('error', (err) => {
+          endTest(false, err.message)
+          resolve()
+        })
+      })
+    }
+
+    // 处理成功的流式响应
+    return new Promise((resolve) => {
+      let buffer = ''
+
+      response.data.on('data', (chunk) => {
+        buffer += chunk.toString()
+        const lines = buffer.split('\n')
+        buffer = lines.pop() || ''
+
+        for (const line of lines) {
+          if (!line.startsWith('data:')) {
+            continue
+          }
+          const jsonStr = line.substring(5).trim()
+          if (!jsonStr || jsonStr === '[DONE]') {
+            continue
+          }
+
+          try {
+            const data = JSON.parse(jsonStr)
+
+            if (data.type === 'content_block_delta' && data.delta?.text) {
+              sendSSE('content', { text: data.delta.text })
+            }
+            if (data.type === 'message_stop') {
+              sendSSE('message_stop')
+            }
+            if (data.type === 'error' || data.error) {
+              const errMsg = data.error?.message || data.message || data.error || 'Unknown error'
+              sendSSE('error', { error: errMsg })
+            }
+          } catch {
+            // ignore parse errors
+          }
+        }
+      })
+
+      response.data.on('end', () => {
+        if (!responseStream.destroyed && !responseStream.writableEnded) {
+          endTest(true)
+        }
+        resolve()
+      })
+
+      response.data.on('error', (err) => {
+        endTest(false, err.message)
+        resolve()
+      })
+    })
+  } catch (error) {
+    logger.error('❌ Stream test request failed:', error.message)
+    endTest(false, error.message)
+  }
+}
+
+module.exports = {
+  randomHex,
+  generateSessionString,
+  createClaudeTestPayload,
+  sendStreamTestRequest
+}

src/utils/unstableUpstreamHelper.js

@@ -0,0 +1,81 @@
+const logger = require('./logger')
+
+function parseList(envValue) {
+  if (!envValue) {
+    return []
+  }
+  return envValue
+    .split(',')
+    .map((s) => s.trim().toLowerCase())
+    .filter(Boolean)
+}
+
+const unstableTypes = new Set(parseList(process.env.UNSTABLE_ERROR_TYPES))
+const unstableKeywords = parseList(process.env.UNSTABLE_ERROR_KEYWORDS)
+const unstableStatusCodes = new Set([408, 499, 502, 503, 504, 522])
+
+function normalizeErrorPayload(payload) {
+  if (!payload) {
+    return {}
+  }
+
+  if (typeof payload === 'string') {
+    try {
+      return normalizeErrorPayload(JSON.parse(payload))
+    } catch (e) {
+      return { message: payload }
+    }
+  }
+
+  if (payload.error && typeof payload.error === 'object') {
+    return {
+      type: payload.error.type || payload.error.error || payload.error.code,
+      code: payload.error.code || payload.error.error || payload.error.type,
+      message: payload.error.message || payload.error.msg || payload.message || payload.error.error
+    }
+  }
+
+  return {
+    type: payload.type || payload.code,
+    code: payload.code || payload.type,
+    message: payload.message || ''
+  }
+}
+
+function isUnstableUpstreamError(statusCode, payload) {
+  const normalizedStatus = Number(statusCode)
+  if (Number.isFinite(normalizedStatus) && normalizedStatus >= 500) {
+    return true
+  }
+  if (Number.isFinite(normalizedStatus) && unstableStatusCodes.has(normalizedStatus)) {
+    return true
+  }
+
+  const { type, code, message } = normalizeErrorPayload(payload)
+  const lowerType = (type || '').toString().toLowerCase()
+  const lowerCode = (code || '').toString().toLowerCase()
+  const lowerMessage = (message || '').toString().toLowerCase()
+
+  if (lowerType === 'server_error' || lowerCode === 'server_error') {
+    return true
+  }
+  if (unstableTypes.has(lowerType) || unstableTypes.has(lowerCode)) {
+    return true
+  }
+  if (unstableKeywords.length > 0) {
+    return unstableKeywords.some((kw) => lowerMessage.includes(kw))
+  }
+
+  return false
+}
+
+function logUnstable(accountLabel, statusCode) {
+  logger.warn(
+    `Detected unstable upstream error (${statusCode}) for account ${accountLabel}, marking temporarily unavailable`
+  )
+}
+
+module.exports = {
+  isUnstableUpstreamError,
+  logUnstable
+}

src/utils/warmupInterceptor.js

@@ -0,0 +1,202 @@
+'use strict'
+
+const { v4: uuidv4 } = require('uuid')
+
+/**
+ * 预热请求拦截器
+ * 检测并拦截低价值请求（标题生成、Warmup等），直接返回模拟响应
+ */
+
+/**
+ * 检测是否为预热请求
+ * @param {Object} body - 请求体
+ * @returns {boolean}
+ */
+function isWarmupRequest(body) {
+  if (!body) {
+    return false
+  }
+
+  // 检查 messages
+  if (body.messages && Array.isArray(body.messages)) {
+    for (const msg of body.messages) {
+      // 处理 content 为数组的情况
+      if (Array.isArray(msg.content)) {
+        for (const content of msg.content) {
+          if (content.type === 'text' && typeof content.text === 'string') {
+            if (isTitleOrWarmupText(content.text)) {
+              return true
+            }
+          }
+        }
+      }
+      // 处理 content 为字符串的情况
+      if (typeof msg.content === 'string') {
+        if (isTitleOrWarmupText(msg.content)) {
+          return true
+        }
+      }
+    }
+  }
+
+  // 检查 system prompt
+  if (body.system) {
+    const systemText = extractSystemText(body.system)
+    if (isTitleExtractionSystemPrompt(systemText)) {
+      return true
+    }
+  }
+
+  return false
+}
+
+/**
+ * 检查文本是否为标题生成或Warmup请求
+ */
+function isTitleOrWarmupText(text) {
+  if (!text) {
+    return false
+  }
+  return (
+    text.includes('Please write a 5-10 word title for the following conversation:') ||
+    text === 'Warmup'
+  )
+}
+
+/**
+ * 检查system prompt是否为标题提取类型
+ */
+function isTitleExtractionSystemPrompt(systemText) {
+  if (!systemText) {
+    return false
+  }
+  return systemText.includes(
+    'nalyze if this message indicates a new conversation topic. If it does, extract a 2-3 word title'
+  )
+}
+
+/**
+ * 从system字段提取文本
+ */
+function extractSystemText(system) {
+  if (typeof system === 'string') {
+    return system
+  }
+  if (Array.isArray(system)) {
+    return system.map((s) => (typeof s === 'object' ? s.text || '' : String(s))).join('')
+  }
+  return ''
+}
+
+/**
+ * 生成模拟的非流式响应
+ * @param {string} model - 模型名称
+ * @returns {Object}
+ */
+function buildMockWarmupResponse(model) {
+  return {
+    id: `msg_warmup_${uuidv4().replace(/-/g, '').slice(0, 20)}`,
+    type: 'message',
+    role: 'assistant',
+    content: [{ type: 'text', text: 'New Conversation' }],
+    model: model || 'claude-3-5-sonnet-20241022',
+    stop_reason: 'end_turn',
+    stop_sequence: null,
+    usage: {
+      input_tokens: 10,
+      output_tokens: 2
+    }
+  }
+}
+
+/**
+ * 发送模拟的流式响应
+ * @param {Object} res - Express response对象
+ * @param {string} model - 模型名称
+ */
+function sendMockWarmupStream(res, model) {
+  const effectiveModel = model || 'claude-3-5-sonnet-20241022'
+  const messageId = `msg_warmup_${uuidv4().replace(/-/g, '').slice(0, 20)}`
+
+  const events = [
+    {
+      event: 'message_start',
+      data: {
+        message: {
+          content: [],
+          id: messageId,
+          model: effectiveModel,
+          role: 'assistant',
+          stop_reason: null,
+          stop_sequence: null,
+          type: 'message',
+          usage: { input_tokens: 10, output_tokens: 0 }
+        },
+        type: 'message_start'
+      }
+    },
+    {
+      event: 'content_block_start',
+      data: {
+        content_block: { text: '', type: 'text' },
+        index: 0,
+        type: 'content_block_start'
+      }
+    },
+    {
+      event: 'content_block_delta',
+      data: {
+        delta: { text: 'New', type: 'text_delta' },
+        index: 0,
+        type: 'content_block_delta'
+      }
+    },
+    {
+      event: 'content_block_delta',
+      data: {
+        delta: { text: ' Conversation', type: 'text_delta' },
+        index: 0,
+        type: 'content_block_delta'
+      }
+    },
+    {
+      event: 'content_block_stop',
+      data: { index: 0, type: 'content_block_stop' }
+    },
+    {
+      event: 'message_delta',
+      data: {
+        delta: { stop_reason: 'end_turn', stop_sequence: null },
+        type: 'message_delta',
+        usage: { input_tokens: 10, output_tokens: 2 }
+      }
+    },
+    {
+      event: 'message_stop',
+      data: { type: 'message_stop' }
+    }
+  ]
+
+  let index = 0
+  const sendNext = () => {
+    if (index >= events.length) {
+      res.end()
+      return
+    }
+
+    const { event, data } = events[index]
+    res.write(`event: ${event}\ndata: ${JSON.stringify(data)}\n\n`)
+    index++
+
+    // 模拟网络延迟
+    setTimeout(sendNext, 20)
+  }
+
+  sendNext()
+}
+
+module.exports = {
+  isWarmupRequest,
+  buildMockWarmupResponse,
+  sendMockWarmupStream
+}

src/validators/clients/claudeCodeValidator.js

@@ -59,13 +59,19 @@ class ClaudeCodeValidator {
       typeof customThreshold === 'number' && Number.isFinite(customThreshold)
         ? customThreshold
         : SYSTEM_PROMPT_THRESHOLD
-
+    
     for (const entry of systemEntries) {
       const rawText = typeof entry?.text === 'string' ? entry.text : ''
-      const { bestScore } = bestSimilarityByTemplates(rawText)
+      const { bestScore, templateId, maskedRaw } = bestSimilarityByTemplates(rawText)
       if (bestScore < threshold) {
         logger.error(
-          `Claude system prompt similarity below threshold: score=${bestScore.toFixed(4)}, threshold=${threshold}, prompt=${rawText}`
+          `Claude system prompt similarity below threshold: score=${bestScore.toFixed(4)}, threshold=${threshold}`
+        )
+        const preview = typeof maskedRaw === 'string' ? maskedRaw.slice(0, 200) : ''
+        logger.warn(
+          `Claude system prompt detail: templateId=${templateId || 'unknown'}, preview=${preview}${
+            maskedRaw && maskedRaw.length > 200 ? '…' : ''
+          }`
         )
         return false
       }

src/validators/clients/codexCliValidator.js

@@ -125,8 +125,12 @@ class CodexCliValidator {
       const part1 = parts1[i] || 0
       const part2 = parts2[i] || 0
 
-      if (part1 < part2) return -1
-      if (part1 > part2) return 1
+      if (part1 < part2) {
+        return -1
+      }
+      if (part1 > part2) {
+        return 1
+      }
     }
 
     return 0

src/validators/clients/geminiCliValidator.js

@@ -53,7 +53,7 @@ class GeminiCliValidator {
       // 2. 对于 /gemini 路径，检查是否包含 generateContent
       if (path.includes('generateContent')) {
         // 包含 generateContent 的路径需要验证 User-Agent
-        const geminiCliPattern = /^GeminiCLI\/v?[\d\.]+/i
+        const geminiCliPattern = /^GeminiCLI\/v?[\d.]+/i
         if (!geminiCliPattern.test(userAgent)) {
           logger.debug(
             `Gemini CLI validation failed - UA mismatch for generateContent: ${userAgent}`
@@ -84,8 +84,12 @@ class GeminiCliValidator {
       const part1 = parts1[i] || 0
       const part2 = parts2[i] || 0
 
-      if (part1 < part2) return -1
-      if (part1 > part2) return 1
+      if (part1 < part2) {
+        return -1
+      }
+      if (part1 > part2) {
+        return 1
+      }
     }
 
     return 0

tests/accountBalanceService.test.js

@@ -0,0 +1,218 @@
+// Mock logger，避免测试输出污染控制台
+jest.mock('../src/utils/logger', () => ({
+  debug: jest.fn(),
+  info: jest.fn(),
+  warn: jest.fn(),
+  error: jest.fn()
+}))
+
+const accountBalanceServiceModule = require('../src/services/accountBalanceService')
+
+const { AccountBalanceService } = accountBalanceServiceModule
+
+describe('AccountBalanceService', () => {
+  const originalBalanceScriptEnabled = process.env.BALANCE_SCRIPT_ENABLED
+
+  afterEach(() => {
+    if (originalBalanceScriptEnabled === undefined) {
+      delete process.env.BALANCE_SCRIPT_ENABLED
+    } else {
+      process.env.BALANCE_SCRIPT_ENABLED = originalBalanceScriptEnabled
+    }
+  })
+
+  const mockLogger = {
+    debug: jest.fn(),
+    info: jest.fn(),
+    warn: jest.fn(),
+    error: jest.fn()
+  }
+
+  const buildMockRedis = () => ({
+    getLocalBalance: jest.fn().mockResolvedValue(null),
+    setLocalBalance: jest.fn().mockResolvedValue(undefined),
+    getAccountBalance: jest.fn().mockResolvedValue(null),
+    setAccountBalance: jest.fn().mockResolvedValue(undefined),
+    deleteAccountBalance: jest.fn().mockResolvedValue(undefined),
+    getBalanceScriptConfig: jest.fn().mockResolvedValue(null),
+    getAccountUsageStats: jest.fn().mockResolvedValue({
+      total: { requests: 10 },
+      daily: { requests: 2, cost: 20 },
+      monthly: { requests: 5 }
+    }),
+    getDateInTimezone: (date) => new Date(date.getTime() + 8 * 3600 * 1000)
+  })
+
+  it('should normalize platform aliases', () => {
+    const service = new AccountBalanceService({ redis: buildMockRedis(), logger: mockLogger })
+    expect(service.normalizePlatform('claude-official')).toBe('claude')
+    expect(service.normalizePlatform('azure-openai')).toBe('azure_openai')
+    expect(service.normalizePlatform('gemini-api')).toBe('gemini-api')
+  })
+
+  it('should build local quota/balance from dailyQuota and local dailyCost', async () => {
+    const mockRedis = buildMockRedis()
+    const service = new AccountBalanceService({ redis: mockRedis, logger: mockLogger })
+
+    service._computeMonthlyCost = jest.fn().mockResolvedValue(30)
+    service._computeTotalCost = jest.fn().mockResolvedValue(123.45)
+
+    const account = { id: 'acct-1', name: 'A', dailyQuota: '100', quotaResetTime: '00:00' }
+    const result = await service._getAccountBalanceForAccount(account, 'claude-console', {
+      queryApi: false,
+      useCache: true
+    })
+
+    expect(result.success).toBe(true)
+    expect(result.data.source).toBe('local')
+    expect(result.data.balance.amount).toBeCloseTo(80, 6)
+    expect(result.data.quota.percentage).toBeCloseTo(20, 6)
+    expect(result.data.statistics.totalCost).toBeCloseTo(123.45, 6)
+    expect(mockRedis.setLocalBalance).toHaveBeenCalled()
+  })
+
+  it('should use cached balance when account has no dailyQuota', async () => {
+    const mockRedis = buildMockRedis()
+    mockRedis.getAccountBalance.mockResolvedValue({
+      status: 'success',
+      balance: 12.34,
+      currency: 'USD',
+      quota: null,
+      errorMessage: '',
+      lastRefreshAt: '2025-01-01T00:00:00Z',
+      ttlSeconds: 120
+    })
+
+    const service = new AccountBalanceService({ redis: mockRedis, logger: mockLogger })
+    service._computeMonthlyCost = jest.fn().mockResolvedValue(0)
+    service._computeTotalCost = jest.fn().mockResolvedValue(0)
+
+    const account = { id: 'acct-2', name: 'B' }
+    const result = await service._getAccountBalanceForAccount(account, 'openai', {
+      queryApi: false,
+      useCache: true
+    })
+
+    expect(result.data.source).toBe('cache')
+    expect(result.data.balance.amount).toBeCloseTo(12.34, 6)
+    expect(result.data.lastRefreshAt).toBe('2025-01-01T00:00:00Z')
+  })
+
+  it('should not cache provider errors and fallback to local when queryApi=true', async () => {
+    const mockRedis = buildMockRedis()
+    const service = new AccountBalanceService({ redis: mockRedis, logger: mockLogger })
+
+    service._computeMonthlyCost = jest.fn().mockResolvedValue(0)
+    service._computeTotalCost = jest.fn().mockResolvedValue(0)
+
+    service.registerProvider('openai', {
+      queryBalance: () => {
+        throw new Error('boom')
+      }
+    })
+
+    const account = { id: 'acct-3', name: 'C' }
+    const result = await service._getAccountBalanceForAccount(account, 'openai', {
+      queryApi: true,
+      useCache: false
+    })
+
+    expect(mockRedis.setAccountBalance).not.toHaveBeenCalled()
+    expect(result.data.source).toBe('local')
+    expect(result.data.status).toBe('error')
+    expect(result.data.error).toBe('boom')
+  })
+
+  it('should ignore script config when balance script is disabled', async () => {
+    process.env.BALANCE_SCRIPT_ENABLED = 'false'
+
+    const mockRedis = buildMockRedis()
+    mockRedis.getBalanceScriptConfig.mockResolvedValue({
+      scriptBody: '({ request: { url: "http://example.com" }, extractor: function(){ return {} } })'
+    })
+
+    const service = new AccountBalanceService({ redis: mockRedis, logger: mockLogger })
+    service._computeMonthlyCost = jest.fn().mockResolvedValue(0)
+    service._computeTotalCost = jest.fn().mockResolvedValue(0)
+
+    const provider = { queryBalance: jest.fn().mockResolvedValue({ balance: 1, currency: 'USD' }) }
+    service.registerProvider('openai', provider)
+
+    const scriptSpy = jest.spyOn(service, '_getBalanceFromScript')
+
+    const account = { id: 'acct-script-off', name: 'S' }
+    const result = await service._getAccountBalanceForAccount(account, 'openai', {
+      queryApi: true,
+      useCache: false
+    })
+
+    expect(provider.queryBalance).toHaveBeenCalled()
+    expect(scriptSpy).not.toHaveBeenCalled()
+    expect(result.data.source).toBe('api')
+  })
+
+  it('should prefer script when configured and enabled', async () => {
+    process.env.BALANCE_SCRIPT_ENABLED = 'true'
+
+    const mockRedis = buildMockRedis()
+    mockRedis.getBalanceScriptConfig.mockResolvedValue({
+      scriptBody: '({ request: { url: "http://example.com" }, extractor: function(){ return {} } })'
+    })
+
+    const service = new AccountBalanceService({ redis: mockRedis, logger: mockLogger })
+    service._computeMonthlyCost = jest.fn().mockResolvedValue(0)
+    service._computeTotalCost = jest.fn().mockResolvedValue(0)
+
+    const provider = { queryBalance: jest.fn().mockResolvedValue({ balance: 2, currency: 'USD' }) }
+    service.registerProvider('openai', provider)
+
+    jest.spyOn(service, '_getBalanceFromScript').mockResolvedValue({
+      status: 'success',
+      balance: 3,
+      currency: 'USD',
+      quota: null,
+      queryMethod: 'script',
+      rawData: { ok: true },
+      lastRefreshAt: '2025-01-01T00:00:00Z',
+      errorMessage: ''
+    })
+
+    const account = { id: 'acct-script-on', name: 'T' }
+    const result = await service._getAccountBalanceForAccount(account, 'openai', {
+      queryApi: true,
+      useCache: false
+    })
+
+    expect(provider.queryBalance).not.toHaveBeenCalled()
+    expect(result.data.source).toBe('api')
+    expect(result.data.balance.amount).toBeCloseTo(3, 6)
+    expect(result.data.lastRefreshAt).toBe('2025-01-01T00:00:00Z')
+  })
+
+  it('should count low balance once per account in summary', async () => {
+    const mockRedis = buildMockRedis()
+    const service = new AccountBalanceService({ redis: mockRedis, logger: mockLogger })
+
+    service.getSupportedPlatforms = () => ['claude-console']
+    service.getAllAccountsByPlatform = async () => [{ id: 'acct-4', name: 'D' }]
+    service._getAccountBalanceForAccount = async () => ({
+      success: true,
+      data: {
+        accountId: 'acct-4',
+        platform: 'claude-console',
+        balance: { amount: 5, currency: 'USD', formattedAmount: '$5.00' },
+        quota: { percentage: 95 },
+        statistics: { totalCost: 1 },
+        source: 'local',
+        lastRefreshAt: '2025-01-01T00:00:00Z',
+        cacheExpiresAt: null,
+        status: 'success',
+        error: null
+      }
+    })
+
+    const summary = await service.getBalanceSummary()
+    expect(summary.lowBalanceCount).toBe(1)
+    expect(summary.platforms['claude-console'].lowBalanceCount).toBe(1)
+  })
+})

tests/concurrencyQueue.integration.test.js

@@ -0,0 +1,860 @@
+/**
+ * 并发请求排队功能集成测试
+ *
+ * 测试分为三个层次：
+ * 1. Mock 测试 - 测试核心逻辑，不需要真实 Redis
+ * 2. Redis 方法测试 - 测试 Redis 操作的原子性和正确性
+ * 3. 端到端场景测试 - 测试完整的排队流程
+ *
+ * 运行方式：
+ * - npm test -- concurrencyQueue.integration  # 运行所有测试（Mock 部分）
+ * - REDIS_TEST=1 npm test -- concurrencyQueue.integration  # 包含真实 Redis 测试
+ */
+
+// Mock logger to avoid console output during tests
+jest.mock('../src/utils/logger', () => ({
+  api: jest.fn(),
+  warn: jest.fn(),
+  error: jest.fn(),
+  info: jest.fn(),
+  database: jest.fn(),
+  security: jest.fn()
+}))
+
+const redis = require('../src/models/redis')
+const claudeRelayConfigService = require('../src/services/claudeRelayConfigService')
+
+// Helper: sleep function
+const sleep = (ms) => new Promise((resolve) => setTimeout(resolve, ms))
+
+// Helper: 创建模拟的 req/res 对象
+function createMockReqRes() {
+  const listeners = {}
+  const req = {
+    destroyed: false,
+    once: jest.fn((event, handler) => {
+      listeners[`req:${event}`] = handler
+    }),
+    removeListener: jest.fn((event) => {
+      delete listeners[`req:${event}`]
+    }),
+    // 触发事件的辅助方法
+    emit: (event) => {
+      const handler = listeners[`req:${event}`]
+      if (handler) {
+        handler()
+      }
+    }
+  }
+
+  const res = {
+    once: jest.fn((event, handler) => {
+      listeners[`res:${event}`] = handler
+    }),
+    removeListener: jest.fn((event) => {
+      delete listeners[`res:${event}`]
+    }),
+    emit: (event) => {
+      const handler = listeners[`res:${event}`]
+      if (handler) {
+        handler()
+      }
+    }
+  }
+
+  return { req, res, listeners }
+}
+
+// ============================================
+// 第一部分：Mock 测试 - waitForConcurrencySlot 核心逻辑
+// ============================================
+describe('ConcurrencyQueue Integration Tests', () => {
+  describe('Part 1: waitForConcurrencySlot Logic (Mocked)', () => {
+    // 导入 auth 模块中的 waitForConcurrencySlot
+    // 由于它是内部函数，我们需要通过测试其行为来验证
+    // 这里我们模拟整个流程
+
+    let mockRedis
+
+    beforeEach(() => {
+      jest.clearAllMocks()
+
+      // 创建 Redis mock
+      mockRedis = {
+        concurrencyCount: {},
+        queueCount: {},
+        stats: {},
+        waitTimes: {},
+        globalWaitTimes: []
+      }
+
+      // Mock Redis 并发方法
+      jest.spyOn(redis, 'incrConcurrency').mockImplementation(async (keyId, requestId, _lease) => {
+        if (!mockRedis.concurrencyCount[keyId]) {
+          mockRedis.concurrencyCount[keyId] = new Set()
+        }
+        mockRedis.concurrencyCount[keyId].add(requestId)
+        return mockRedis.concurrencyCount[keyId].size
+      })
+
+      jest.spyOn(redis, 'decrConcurrency').mockImplementation(async (keyId, requestId) => {
+        if (mockRedis.concurrencyCount[keyId]) {
+          mockRedis.concurrencyCount[keyId].delete(requestId)
+          return mockRedis.concurrencyCount[keyId].size
+        }
+        return 0
+      })
+
+      // Mock 排队计数方法
+      jest.spyOn(redis, 'incrConcurrencyQueue').mockImplementation(async (keyId) => {
+        mockRedis.queueCount[keyId] = (mockRedis.queueCount[keyId] || 0) + 1
+        return mockRedis.queueCount[keyId]
+      })
+
+      jest.spyOn(redis, 'decrConcurrencyQueue').mockImplementation(async (keyId) => {
+        mockRedis.queueCount[keyId] = Math.max(0, (mockRedis.queueCount[keyId] || 0) - 1)
+        return mockRedis.queueCount[keyId]
+      })
+
+      jest
+        .spyOn(redis, 'getConcurrencyQueueCount')
+        .mockImplementation(async (keyId) => mockRedis.queueCount[keyId] || 0)
+
+      // Mock 统计方法
+      jest.spyOn(redis, 'incrConcurrencyQueueStats').mockImplementation(async (keyId, field) => {
+        if (!mockRedis.stats[keyId]) {
+          mockRedis.stats[keyId] = {}
+        }
+        mockRedis.stats[keyId][field] = (mockRedis.stats[keyId][field] || 0) + 1
+        return mockRedis.stats[keyId][field]
+      })
+
+      jest.spyOn(redis, 'recordQueueWaitTime').mockResolvedValue(undefined)
+      jest.spyOn(redis, 'recordGlobalQueueWaitTime').mockResolvedValue(undefined)
+    })
+
+    afterEach(() => {
+      jest.restoreAllMocks()
+    })
+
+    describe('Slot Acquisition Flow', () => {
+      it('should acquire slot immediately when under concurrency limit', async () => {
+        // 模拟 waitForConcurrencySlot 的行为
+        const keyId = 'test-key-1'
+        const requestId = 'req-1'
+        const concurrencyLimit = 5
+
+        // 直接测试 incrConcurrency 的行为
+        const count = await redis.incrConcurrency(keyId, requestId, 300)
+
+        expect(count).toBe(1)
+        expect(count).toBeLessThanOrEqual(concurrencyLimit)
+      })
+
+      it('should track multiple concurrent requests correctly', async () => {
+        const keyId = 'test-key-2'
+        const concurrencyLimit = 3
+
+        // 模拟多个并发请求
+        const results = []
+        for (let i = 1; i <= 5; i++) {
+          const count = await redis.incrConcurrency(keyId, `req-${i}`, 300)
+          results.push({ requestId: `req-${i}`, count, exceeds: count > concurrencyLimit })
+        }
+
+        // 前3个应该在限制内
+        expect(results[0].exceeds).toBe(false)
+        expect(results[1].exceeds).toBe(false)
+        expect(results[2].exceeds).toBe(false)
+        // 后2个超过限制
+        expect(results[3].exceeds).toBe(true)
+        expect(results[4].exceeds).toBe(true)
+      })
+
+      it('should release slot and allow next request', async () => {
+        const keyId = 'test-key-3'
+        const concurrencyLimit = 1
+
+        // 第一个请求获取槽位
+        const count1 = await redis.incrConcurrency(keyId, 'req-1', 300)
+        expect(count1).toBe(1)
+
+        // 第二个请求超限
+        const count2 = await redis.incrConcurrency(keyId, 'req-2', 300)
+        expect(count2).toBe(2)
+        expect(count2).toBeGreaterThan(concurrencyLimit)
+
+        // 释放第二个请求（因为超限）
+        await redis.decrConcurrency(keyId, 'req-2')
+
+        // 释放第一个请求
+        await redis.decrConcurrency(keyId, 'req-1')
+
+        // 现在第三个请求应该能获取
+        const count3 = await redis.incrConcurrency(keyId, 'req-3', 300)
+        expect(count3).toBe(1)
+      })
+    })
+
+    describe('Queue Count Management', () => {
+      it('should increment and decrement queue count atomically', async () => {
+        const keyId = 'test-key-4'
+
+        // 增加排队计数
+        const count1 = await redis.incrConcurrencyQueue(keyId, 60000)
+        expect(count1).toBe(1)
+
+        const count2 = await redis.incrConcurrencyQueue(keyId, 60000)
+        expect(count2).toBe(2)
+
+        // 减少排队计数
+        const count3 = await redis.decrConcurrencyQueue(keyId)
+        expect(count3).toBe(1)
+
+        const count4 = await redis.decrConcurrencyQueue(keyId)
+        expect(count4).toBe(0)
+      })
+
+      it('should not go below zero on decrement', async () => {
+        const keyId = 'test-key-5'
+
+        // 直接减少（没有先增加）
+        const count = await redis.decrConcurrencyQueue(keyId)
+        expect(count).toBe(0)
+      })
+
+      it('should handle concurrent queue operations', async () => {
+        const keyId = 'test-key-6'
+
+        // 并发增加
+        const increments = await Promise.all([
+          redis.incrConcurrencyQueue(keyId, 60000),
+          redis.incrConcurrencyQueue(keyId, 60000),
+          redis.incrConcurrencyQueue(keyId, 60000)
+        ])
+
+        // 所有增量应该是连续的
+        const sortedIncrements = [...increments].sort((a, b) => a - b)
+        expect(sortedIncrements).toEqual([1, 2, 3])
+      })
+    })
+
+    describe('Statistics Tracking', () => {
+      it('should track entered/success/timeout/cancelled stats', async () => {
+        const keyId = 'test-key-7'
+
+        await redis.incrConcurrencyQueueStats(keyId, 'entered')
+        await redis.incrConcurrencyQueueStats(keyId, 'entered')
+        await redis.incrConcurrencyQueueStats(keyId, 'success')
+        await redis.incrConcurrencyQueueStats(keyId, 'timeout')
+        await redis.incrConcurrencyQueueStats(keyId, 'cancelled')
+
+        expect(mockRedis.stats[keyId]).toEqual({
+          entered: 2,
+          success: 1,
+          timeout: 1,
+          cancelled: 1
+        })
+      })
+    })
+
+    describe('Client Disconnection Handling', () => {
+      it('should detect client disconnection via close event', async () => {
+        const { req } = createMockReqRes()
+
+        let clientDisconnected = false
+
+        // 设置监听器
+        req.once('close', () => {
+          clientDisconnected = true
+        })
+
+        // 模拟客户端断开
+        req.emit('close')
+
+        expect(clientDisconnected).toBe(true)
+      })
+
+      it('should detect pre-destroyed request', () => {
+        const { req } = createMockReqRes()
+        req.destroyed = true
+
+        expect(req.destroyed).toBe(true)
+      })
+    })
+
+    describe('Exponential Backoff Simulation', () => {
+      it('should increase poll interval with backoff', () => {
+        const config = {
+          pollIntervalMs: 200,
+          maxPollIntervalMs: 2000,
+          backoffFactor: 1.5,
+          jitterRatio: 0 // 禁用抖动以便测试
+        }
+
+        let interval = config.pollIntervalMs
+        const intervals = [interval]
+
+        for (let i = 0; i < 5; i++) {
+          interval = Math.min(interval * config.backoffFactor, config.maxPollIntervalMs)
+          intervals.push(interval)
+        }
+
+        // 验证指数增长
+        expect(intervals[1]).toBe(300) // 200 * 1.5
+        expect(intervals[2]).toBe(450) // 300 * 1.5
+        expect(intervals[3]).toBe(675) // 450 * 1.5
+        expect(intervals[4]).toBe(1012.5) // 675 * 1.5
+        expect(intervals[5]).toBe(1518.75) // 1012.5 * 1.5
+      })
+
+      it('should cap interval at maximum', () => {
+        const config = {
+          pollIntervalMs: 1000,
+          maxPollIntervalMs: 2000,
+          backoffFactor: 1.5
+        }
+
+        let interval = config.pollIntervalMs
+
+        for (let i = 0; i < 10; i++) {
+          interval = Math.min(interval * config.backoffFactor, config.maxPollIntervalMs)
+        }
+
+        expect(interval).toBe(2000)
+      })
+
+      it('should apply jitter within expected range', () => {
+        const baseInterval = 1000
+        const jitterRatio = 0.2 // ±20%
+        const results = []
+
+        for (let i = 0; i < 100; i++) {
+          const randomValue = Math.random()
+          const jitter = baseInterval * jitterRatio * (randomValue * 2 - 1)
+          const finalInterval = baseInterval + jitter
+          results.push(finalInterval)
+        }
+
+        const min = Math.min(...results)
+        const max = Math.max(...results)
+
+        // 所有结果应该在 [800, 1200] 范围内
+        expect(min).toBeGreaterThanOrEqual(800)
+        expect(max).toBeLessThanOrEqual(1200)
+      })
+    })
+  })
+
+  // ============================================
+  // 第二部分：并发竞争场景测试
+  // ============================================
+  describe('Part 2: Concurrent Race Condition Tests', () => {
+    beforeEach(() => {
+      jest.clearAllMocks()
+    })
+
+    afterEach(() => {
+      jest.restoreAllMocks()
+    })
+
+    describe('Race Condition: Multiple Requests Competing for Same Slot', () => {
+      it('should handle race condition when multiple requests try to acquire last slot', async () => {
+        const keyId = 'race-test-1'
+        const concurrencyLimit = 1
+        const concurrencyState = { count: 0, holders: new Set() }
+
+        // 模拟原子的 incrConcurrency
+        jest.spyOn(redis, 'incrConcurrency').mockImplementation(async (key, reqId) => {
+          // 模拟原子操作
+          concurrencyState.count++
+          concurrencyState.holders.add(reqId)
+          return concurrencyState.count
+        })
+
+        jest.spyOn(redis, 'decrConcurrency').mockImplementation(async (key, reqId) => {
+          if (concurrencyState.holders.has(reqId)) {
+            concurrencyState.count--
+            concurrencyState.holders.delete(reqId)
+          }
+          return concurrencyState.count
+        })
+
+        // 5个请求同时竞争1个槽位
+        const requests = Array.from({ length: 5 }, (_, i) => `req-${i + 1}`)
+
+        const acquireResults = await Promise.all(
+          requests.map(async (reqId) => {
+            const count = await redis.incrConcurrency(keyId, reqId, 300)
+            const acquired = count <= concurrencyLimit
+
+            if (!acquired) {
+              // 超限，释放
+              await redis.decrConcurrency(keyId, reqId)
+            }
+
+            return { reqId, count, acquired }
+          })
+        )
+
+        // 只有一个请求应该成功获取槽位
+        const successfulAcquires = acquireResults.filter((r) => r.acquired)
+        expect(successfulAcquires.length).toBe(1)
+
+        // 最终并发计数应该是1
+        expect(concurrencyState.count).toBe(1)
+      })
+
+      it('should maintain consistency under high contention', async () => {
+        const keyId = 'race-test-2'
+        const concurrencyLimit = 3
+        const requestCount = 20
+        const concurrencyState = { count: 0, maxSeen: 0 }
+
+        jest.spyOn(redis, 'incrConcurrency').mockImplementation(async () => {
+          concurrencyState.count++
+          concurrencyState.maxSeen = Math.max(concurrencyState.maxSeen, concurrencyState.count)
+          return concurrencyState.count
+        })
+
+        jest.spyOn(redis, 'decrConcurrency').mockImplementation(async () => {
+          concurrencyState.count = Math.max(0, concurrencyState.count - 1)
+          return concurrencyState.count
+        })
+
+        // 模拟多轮请求
+        const activeRequests = []
+
+        for (let i = 0; i < requestCount; i++) {
+          const count = await redis.incrConcurrency(keyId, `req-${i}`, 300)
+
+          if (count <= concurrencyLimit) {
+            activeRequests.push(`req-${i}`)
+
+            // 模拟处理时间后释放
+            setTimeout(async () => {
+              await redis.decrConcurrency(keyId, `req-${i}`)
+            }, Math.random() * 50)
+          } else {
+            await redis.decrConcurrency(keyId, `req-${i}`)
+          }
+
+          // 随机延迟
+          await sleep(Math.random() * 10)
+        }
+
+        // 等待所有请求完成
+        await sleep(100)
+
+        // 最大并发不应超过限制
+        expect(concurrencyState.maxSeen).toBeLessThanOrEqual(concurrencyLimit + requestCount) // 允许短暂超限
+      })
+    })
+
+    describe('Queue Overflow Protection', () => {
+      it('should reject requests when queue is full', async () => {
+        const keyId = 'overflow-test-1'
+        const maxQueueSize = 5
+        const queueState = { count: 0 }
+
+        jest.spyOn(redis, 'incrConcurrencyQueue').mockImplementation(async () => {
+          queueState.count++
+          return queueState.count
+        })
+
+        jest.spyOn(redis, 'decrConcurrencyQueue').mockImplementation(async () => {
+          queueState.count = Math.max(0, queueState.count - 1)
+          return queueState.count
+        })
+
+        const results = []
+
+        // 尝试10个请求进入队列
+        for (let i = 0; i < 10; i++) {
+          const queueCount = await redis.incrConcurrencyQueue(keyId, 60000)
+
+          if (queueCount > maxQueueSize) {
+            // 队列满，释放并拒绝
+            await redis.decrConcurrencyQueue(keyId)
+            results.push({ index: i, accepted: false })
+          } else {
+            results.push({ index: i, accepted: true, position: queueCount })
+          }
+        }
+
+        const accepted = results.filter((r) => r.accepted)
+        const rejected = results.filter((r) => !r.accepted)
+
+        expect(accepted.length).toBe(5)
+        expect(rejected.length).toBe(5)
+      })
+    })
+  })
+
+  // ============================================
+  // 第三部分：真实 Redis 集成测试（可选）
+  // ============================================
+  describe('Part 3: Real Redis Integration Tests', () => {
+    const skipRealRedis = !process.env.REDIS_TEST
+
+    // 辅助函数：检查 Redis 连接
+    async function checkRedisConnection() {
+      try {
+        const client = redis.getClient()
+        if (!client) {
+          return false
+        }
+        await client.ping()
+        return true
+      } catch {
+        return false
+      }
+    }
+
+    beforeAll(async () => {
+      if (skipRealRedis) {
+        console.log('⏭️  Skipping real Redis tests (set REDIS_TEST=1 to enable)')
+        return
+      }
+
+      const connected = await checkRedisConnection()
+      if (!connected) {
+        console.log('⚠️  Redis not connected, skipping real Redis tests')
+      }
+    })
+
+    // 清理测试数据
+    afterEach(async () => {
+      if (skipRealRedis) {
+        return
+      }
+
+      try {
+        const client = redis.getClient()
+        if (!client) {
+          return
+        }
+
+        // 清理测试键
+        const testKeys = await client.keys('concurrency:queue:test-*')
+        if (testKeys.length > 0) {
+          await client.del(...testKeys)
+        }
+      } catch {
+        // 忽略清理错误
+      }
+    })
+
+    describe('Redis Queue Operations', () => {
+      const testOrSkip = skipRealRedis ? it.skip : it
+
+      testOrSkip('should atomically increment queue count with TTL', async () => {
+        const keyId = 'test-redis-queue-1'
+        const timeoutMs = 5000
+
+        const count1 = await redis.incrConcurrencyQueue(keyId, timeoutMs)
+        expect(count1).toBe(1)
+
+        const count2 = await redis.incrConcurrencyQueue(keyId, timeoutMs)
+        expect(count2).toBe(2)
+
+        // 验证 TTL 被设置
+        const client = redis.getClient()
+        const ttl = await client.ttl(`concurrency:queue:${keyId}`)
+        expect(ttl).toBeGreaterThan(0)
+        expect(ttl).toBeLessThanOrEqual(Math.ceil(timeoutMs / 1000) + 30)
+      })
+
+      testOrSkip('should atomically decrement and delete when zero', async () => {
+        const keyId = 'test-redis-queue-2'
+
+        await redis.incrConcurrencyQueue(keyId, 60000)
+        const count = await redis.decrConcurrencyQueue(keyId)
+
+        expect(count).toBe(0)
+
+        // 验证键已删除
+        const client = redis.getClient()
+        const exists = await client.exists(`concurrency:queue:${keyId}`)
+        expect(exists).toBe(0)
+      })
+
+      testOrSkip('should handle concurrent increments correctly', async () => {
+        const keyId = 'test-redis-queue-3'
+        const numRequests = 10
+
+        // 并发增加
+        const results = await Promise.all(
+          Array.from({ length: numRequests }, () => redis.incrConcurrencyQueue(keyId, 60000))
+        )
+
+        // 所有结果应该是 1 到 numRequests
+        const sorted = [...results].sort((a, b) => a - b)
+        expect(sorted).toEqual(Array.from({ length: numRequests }, (_, i) => i + 1))
+      })
+    })
+
+    describe('Redis Stats Operations', () => {
+      const testOrSkip = skipRealRedis ? it.skip : it
+
+      testOrSkip('should track queue statistics correctly', async () => {
+        const keyId = 'test-redis-stats-1'
+
+        await redis.incrConcurrencyQueueStats(keyId, 'entered')
+        await redis.incrConcurrencyQueueStats(keyId, 'entered')
+        await redis.incrConcurrencyQueueStats(keyId, 'success')
+        await redis.incrConcurrencyQueueStats(keyId, 'timeout')
+
+        const stats = await redis.getConcurrencyQueueStats(keyId)
+
+        expect(stats.entered).toBe(2)
+        expect(stats.success).toBe(1)
+        expect(stats.timeout).toBe(1)
+        expect(stats.cancelled).toBe(0)
+      })
+
+      testOrSkip('should record and retrieve wait times', async () => {
+        const keyId = 'test-redis-wait-1'
+        const waitTimes = [100, 200, 150, 300, 250]
+
+        for (const wt of waitTimes) {
+          await redis.recordQueueWaitTime(keyId, wt)
+        }
+
+        const recorded = await redis.getQueueWaitTimes(keyId)
+
+        // 应该按 LIFO 顺序存储
+        expect(recorded.length).toBe(5)
+        expect(recorded[0]).toBe(250) // 最后插入的在前面
+      })
+
+      testOrSkip('should record global wait times', async () => {
+        const waitTimes = [500, 600, 700]
+
+        for (const wt of waitTimes) {
+          await redis.recordGlobalQueueWaitTime(wt)
+        }
+
+        const recorded = await redis.getGlobalQueueWaitTimes()
+
+        expect(recorded.length).toBeGreaterThanOrEqual(3)
+      })
+    })
+
+    describe('Redis Cleanup Operations', () => {
+      const testOrSkip = skipRealRedis ? it.skip : it
+
+      testOrSkip('should clear specific queue', async () => {
+        const keyId = 'test-redis-clear-1'
+
+        await redis.incrConcurrencyQueue(keyId, 60000)
+        await redis.incrConcurrencyQueue(keyId, 60000)
+
+        const cleared = await redis.clearConcurrencyQueue(keyId)
+        expect(cleared).toBe(true)
+
+        const count = await redis.getConcurrencyQueueCount(keyId)
+        expect(count).toBe(0)
+      })
+
+      testOrSkip('should clear all queues but preserve stats', async () => {
+        const keyId1 = 'test-redis-clearall-1'
+        const keyId2 = 'test-redis-clearall-2'
+
+        // 创建队列和统计
+        await redis.incrConcurrencyQueue(keyId1, 60000)
+        await redis.incrConcurrencyQueue(keyId2, 60000)
+        await redis.incrConcurrencyQueueStats(keyId1, 'entered')
+
+        // 清理所有队列
+        const cleared = await redis.clearAllConcurrencyQueues()
+        expect(cleared).toBeGreaterThanOrEqual(2)
+
+        // 验证队列已清理
+        const count1 = await redis.getConcurrencyQueueCount(keyId1)
+        const count2 = await redis.getConcurrencyQueueCount(keyId2)
+        expect(count1).toBe(0)
+        expect(count2).toBe(0)
+
+        // 统计应该保留
+        const stats = await redis.getConcurrencyQueueStats(keyId1)
+        expect(stats.entered).toBe(1)
+      })
+    })
+  })
+
+  // ============================================
+  // 第四部分：配置服务集成测试
+  // ============================================
+  describe('Part 4: Configuration Service Integration', () => {
+    beforeEach(() => {
+      // 清除配置缓存
+      claudeRelayConfigService.clearCache()
+    })
+
+    afterEach(() => {
+      jest.restoreAllMocks()
+    })
+
+    describe('Queue Configuration', () => {
+      it('should return default queue configuration', async () => {
+        jest.spyOn(redis, 'getClient').mockReturnValue(null)
+
+        const config = await claudeRelayConfigService.getConfig()
+
+        expect(config.concurrentRequestQueueEnabled).toBe(false)
+        expect(config.concurrentRequestQueueMaxSize).toBe(3)
+        expect(config.concurrentRequestQueueMaxSizeMultiplier).toBe(0)
+        expect(config.concurrentRequestQueueTimeoutMs).toBe(10000)
+      })
+
+      it('should calculate max queue size correctly', async () => {
+        const testCases = [
+          { concurrencyLimit: 5, multiplier: 2, fixedMin: 3, expected: 10 }, // 5*2=10 > 3
+          { concurrencyLimit: 1, multiplier: 1, fixedMin: 5, expected: 5 }, // 1*1=1 < 5
+          { concurrencyLimit: 10, multiplier: 0.5, fixedMin: 3, expected: 5 }, // 10*0.5=5 > 3
+          { concurrencyLimit: 2, multiplier: 1, fixedMin: 10, expected: 10 } // 2*1=2 < 10
+        ]
+
+        for (const tc of testCases) {
+          const maxQueueSize = Math.max(tc.concurrencyLimit * tc.multiplier, tc.fixedMin)
+          expect(maxQueueSize).toBe(tc.expected)
+        }
+      })
+    })
+  })
+
+  // ============================================
+  // 第五部分：端到端场景测试
+  // ============================================
+  describe('Part 5: End-to-End Scenario Tests', () => {
+    describe('Scenario: Claude Code Agent Parallel Tool Calls', () => {
+      it('should handle burst of parallel tool results', async () => {
+        // 模拟 Claude Code Agent 发送多个并行工具结果的场景
+        const concurrencyLimit = 2
+        const maxQueueSize = 5
+
+        const state = {
+          concurrency: 0,
+          queue: 0,
+          completed: 0,
+          rejected: 0
+        }
+
+        // 模拟 8 个并行工具结果请求
+        const requests = Array.from({ length: 8 }, (_, i) => ({
+          id: `tool-result-${i + 1}`,
+          startTime: Date.now()
+        }))
+
+        // 模拟处理逻辑
+        async function processRequest(req) {
+          // 尝试获取并发槽位
+          state.concurrency++
+
+          if (state.concurrency > concurrencyLimit) {
+            // 超限，进入队列
+            state.concurrency--
+            state.queue++
+
+            if (state.queue > maxQueueSize) {
+              // 队列满，拒绝
+              state.queue--
+              state.rejected++
+              return { ...req, status: 'rejected', reason: 'queue_full' }
+            }
+
+            // 等待槽位（模拟）
+            await sleep(Math.random() * 100)
+            state.queue--
+            state.concurrency++
+          }
+
+          // 处理请求
+          await sleep(50) // 模拟处理时间
+          state.concurrency--
+          state.completed++
+
+          return { ...req, status: 'completed', duration: Date.now() - req.startTime }
+        }
+
+        const results = await Promise.all(requests.map(processRequest))
+
+        const completed = results.filter((r) => r.status === 'completed')
+        const rejected = results.filter((r) => r.status === 'rejected')
+
+        // 大部分请求应该完成
+        expect(completed.length).toBeGreaterThan(0)
+        // 可能有一些被拒绝
+        expect(state.rejected).toBe(rejected.length)
+
+        console.log(
+          `  ✓ Completed: ${completed.length}, Rejected: ${rejected.length}, Max concurrent: ${concurrencyLimit}`
+        )
+      })
+    })
+
+    describe('Scenario: Graceful Degradation', () => {
+      it('should fallback when Redis fails', async () => {
+        jest
+          .spyOn(redis, 'incrConcurrencyQueue')
+          .mockRejectedValue(new Error('Redis connection lost'))
+
+        // 模拟降级行为：Redis 失败时直接拒绝而不是崩溃
+        let result
+        try {
+          await redis.incrConcurrencyQueue('fallback-test', 60000)
+          result = { success: true }
+        } catch (error) {
+          // 优雅降级：返回 429 而不是 500
+          result = { success: false, fallback: true, error: error.message }
+        }
+
+        expect(result.fallback).toBe(true)
+        expect(result.error).toContain('Redis')
+      })
+    })
+
+    describe('Scenario: Timeout Behavior', () => {
+      it('should respect queue timeout', async () => {
+        const timeoutMs = 100
+        const startTime = Date.now()
+
+        // 模拟等待超时
+        await new Promise((resolve) => setTimeout(resolve, timeoutMs))
+
+        const elapsed = Date.now() - startTime
+        expect(elapsed).toBeGreaterThanOrEqual(timeoutMs - 10) // 允许 10ms 误差
+      })
+
+      it('should track timeout statistics', async () => {
+        const stats = { entered: 0, success: 0, timeout: 0, cancelled: 0 }
+
+        // 模拟多个请求，部分超时
+        const requests = [
+          { id: 'req-1', willTimeout: false },
+          { id: 'req-2', willTimeout: true },
+          { id: 'req-3', willTimeout: false },
+          { id: 'req-4', willTimeout: true }
+        ]
+
+        for (const req of requests) {
+          stats.entered++
+          if (req.willTimeout) {
+            stats.timeout++
+          } else {
+            stats.success++
+          }
+        }
+
+        expect(stats.entered).toBe(4)
+        expect(stats.success).toBe(2)
+        expect(stats.timeout).toBe(2)
+
+        // 成功率应该是 50%
+        const successRate = (stats.success / stats.entered) * 100
+        expect(successRate).toBe(50)
+      })
+    })
+  })
+})

tests/concurrencyQueue.test.js

@@ -0,0 +1,278 @@
+/**
+ * 并发请求排队功能测试
+ * 测试排队逻辑中的核心算法：百分位数计算、等待时间统计、指数退避等
+ *
+ * 注意：Redis 方法的测试需要集成测试环境，这里主要测试纯算法逻辑
+ */
+
+// Mock logger to avoid console output during tests
+jest.mock('../src/utils/logger', () => ({
+  api: jest.fn(),
+  warn: jest.fn(),
+  error: jest.fn(),
+  info: jest.fn(),
+  database: jest.fn(),
+  security: jest.fn()
+}))
+
+// 使用共享的统计工具函数（与生产代码一致）
+const { getPercentile, calculateWaitTimeStats } = require('../src/utils/statsHelper')
+
+describe('ConcurrencyQueue', () => {
+  describe('Percentile Calculation (nearest-rank method)', () => {
+    // 直接测试共享工具函数，确保与生产代码行为一致
+    it('should return 0 for empty array', () => {
+      expect(getPercentile([], 50)).toBe(0)
+    })
+
+    it('should return single element for single-element array', () => {
+      expect(getPercentile([100], 50)).toBe(100)
+      expect(getPercentile([100], 99)).toBe(100)
+    })
+
+    it('should return min for percentile 0', () => {
+      expect(getPercentile([10, 20, 30, 40, 50], 0)).toBe(10)
+    })
+
+    it('should return max for percentile 100', () => {
+      expect(getPercentile([10, 20, 30, 40, 50], 100)).toBe(50)
+    })
+
+    it('should calculate P50 correctly for len=10', () => {
+      // For [10, 20, 30, 40, 50, 60, 70, 80, 90, 100] (len=10)
+      // P50: ceil(50/100 * 10) - 1 = ceil(5) - 1 = 4 → value at index 4 = 50
+      const arr = [10, 20, 30, 40, 50, 60, 70, 80, 90, 100]
+      expect(getPercentile(arr, 50)).toBe(50)
+    })
+
+    it('should calculate P90 correctly for len=10', () => {
+      // For len=10, P90: ceil(90/100 * 10) - 1 = ceil(9) - 1 = 8 → value at index 8 = 90
+      const arr = [10, 20, 30, 40, 50, 60, 70, 80, 90, 100]
+      expect(getPercentile(arr, 90)).toBe(90)
+    })
+
+    it('should calculate P99 correctly for len=100', () => {
+      // For len=100, P99: ceil(99/100 * 100) - 1 = ceil(99) - 1 = 98
+      const arr = Array.from({ length: 100 }, (_, i) => i + 1)
+      expect(getPercentile(arr, 99)).toBe(99)
+    })
+
+    it('should handle two-element array correctly', () => {
+      // For [10, 20] (len=2)
+      // P50: ceil(50/100 * 2) - 1 = ceil(1) - 1 = 0 → value = 10
+      expect(getPercentile([10, 20], 50)).toBe(10)
+    })
+
+    it('should handle negative percentile as 0', () => {
+      expect(getPercentile([10, 20, 30], -10)).toBe(10)
+    })
+
+    it('should handle percentile > 100 as 100', () => {
+      expect(getPercentile([10, 20, 30], 150)).toBe(30)
+    })
+  })
+
+  describe('Wait Time Stats Calculation', () => {
+    // 直接测试共享工具函数
+    it('should return null for empty array', () => {
+      expect(calculateWaitTimeStats([])).toBeNull()
+    })
+
+    it('should return null for null input', () => {
+      expect(calculateWaitTimeStats(null)).toBeNull()
+    })
+
+    it('should return null for undefined input', () => {
+      expect(calculateWaitTimeStats(undefined)).toBeNull()
+    })
+
+    it('should calculate stats correctly for typical data', () => {
+      const waitTimes = [100, 200, 150, 300, 250, 180, 220, 280, 190, 210]
+      const stats = calculateWaitTimeStats(waitTimes)
+
+      expect(stats.count).toBe(10)
+      expect(stats.min).toBe(100)
+      expect(stats.max).toBe(300)
+      // Sum: 100+150+180+190+200+210+220+250+280+300 = 2080
+      expect(stats.avg).toBe(208)
+      expect(stats.sampleSizeWarning).toBeUndefined()
+    })
+
+    it('should add warning for small sample size (< 10)', () => {
+      const waitTimes = [100, 200, 300]
+      const stats = calculateWaitTimeStats(waitTimes)
+
+      expect(stats.count).toBe(3)
+      expect(stats.sampleSizeWarning).toBe('Results may be inaccurate due to small sample size')
+    })
+
+    it('should handle single value', () => {
+      const stats = calculateWaitTimeStats([500])
+
+      expect(stats.count).toBe(1)
+      expect(stats.min).toBe(500)
+      expect(stats.max).toBe(500)
+      expect(stats.avg).toBe(500)
+      expect(stats.p50).toBe(500)
+      expect(stats.p90).toBe(500)
+      expect(stats.p99).toBe(500)
+    })
+
+    it('should sort input array before calculating', () => {
+      const waitTimes = [500, 100, 300, 200, 400]
+      const stats = calculateWaitTimeStats(waitTimes)
+
+      expect(stats.min).toBe(100)
+      expect(stats.max).toBe(500)
+    })
+
+    it('should not modify original array', () => {
+      const waitTimes = [500, 100, 300]
+      calculateWaitTimeStats(waitTimes)
+
+      expect(waitTimes).toEqual([500, 100, 300])
+    })
+  })
+
+  describe('Exponential Backoff with Jitter', () => {
+    /**
+     * 指数退避计算函数（与 auth.js 中的实现一致）
+     * @param {number} currentInterval - 当前轮询间隔
+     * @param {number} backoffFactor - 退避系数
+     * @param {number} jitterRatio - 抖动比例
+     * @param {number} maxInterval - 最大间隔
+     * @param {number} randomValue - 随机值 [0, 1)，用于确定性测试
+     */
+    function calculateNextInterval(
+      currentInterval,
+      backoffFactor,
+      jitterRatio,
+      maxInterval,
+      randomValue
+    ) {
+      let nextInterval = currentInterval * backoffFactor
+      // 抖动范围：[-jitterRatio, +jitterRatio]
+      const jitter = nextInterval * jitterRatio * (randomValue * 2 - 1)
+      nextInterval = nextInterval + jitter
+      return Math.max(1, Math.min(nextInterval, maxInterval))
+    }
+
+    it('should apply exponential backoff without jitter (randomValue=0.5)', () => {
+      // randomValue = 0.5 gives jitter = 0
+      const next = calculateNextInterval(100, 1.5, 0.2, 1000, 0.5)
+      expect(next).toBe(150) // 100 * 1.5 = 150
+    })
+
+    it('should apply maximum positive jitter (randomValue=1.0)', () => {
+      // randomValue = 1.0 gives maximum positive jitter (+20%)
+      const next = calculateNextInterval(100, 1.5, 0.2, 1000, 1.0)
+      // 100 * 1.5 = 150, jitter = 150 * 0.2 * 1 = 30
+      expect(next).toBe(180) // 150 + 30
+    })
+
+    it('should apply maximum negative jitter (randomValue=0.0)', () => {
+      // randomValue = 0.0 gives maximum negative jitter (-20%)
+      const next = calculateNextInterval(100, 1.5, 0.2, 1000, 0.0)
+      // 100 * 1.5 = 150, jitter = 150 * 0.2 * -1 = -30
+      expect(next).toBe(120) // 150 - 30
+    })
+
+    it('should respect maximum interval', () => {
+      const next = calculateNextInterval(800, 1.5, 0.2, 1000, 1.0)
+      // 800 * 1.5 = 1200, with +20% jitter = 1440, capped at 1000
+      expect(next).toBe(1000)
+    })
+
+    it('should never go below 1ms even with extreme negative jitter', () => {
+      const next = calculateNextInterval(1, 1.0, 0.9, 1000, 0.0)
+      // 1 * 1.0 = 1, jitter = 1 * 0.9 * -1 = -0.9
+      // 1 - 0.9 = 0.1, but Math.max(1, ...) ensures minimum is 1
+      expect(next).toBe(1)
+    })
+
+    it('should handle zero jitter ratio', () => {
+      const next = calculateNextInterval(100, 2.0, 0, 1000, 0.0)
+      expect(next).toBe(200) // Pure exponential, no jitter
+    })
+
+    it('should handle large backoff factor', () => {
+      const next = calculateNextInterval(100, 3.0, 0.1, 1000, 0.5)
+      expect(next).toBe(300) // 100 * 3.0 = 300
+    })
+
+    describe('jitter distribution', () => {
+      it('should produce values in expected range', () => {
+        const results = []
+        // Test with various random values
+        for (let r = 0; r <= 1; r += 0.1) {
+          results.push(calculateNextInterval(100, 1.5, 0.2, 1000, r))
+        }
+        // All values should be between 120 (150 - 30) and 180 (150 + 30)
+        expect(Math.min(...results)).toBeGreaterThanOrEqual(120)
+        expect(Math.max(...results)).toBeLessThanOrEqual(180)
+      })
+    })
+  })
+
+  describe('Queue Size Calculation', () => {
+    /**
+     * 最大排队数计算（与 auth.js 中的实现一致）
+     */
+    function calculateMaxQueueSize(concurrencyLimit, multiplier, fixedMin) {
+      return Math.max(concurrencyLimit * multiplier, fixedMin)
+    }
+
+    it('should use multiplier when result is larger', () => {
+      // concurrencyLimit=10, multiplier=2, fixedMin=5
+      // max(10*2, 5) = max(20, 5) = 20
+      expect(calculateMaxQueueSize(10, 2, 5)).toBe(20)
+    })
+
+    it('should use fixed minimum when multiplier result is smaller', () => {
+      // concurrencyLimit=2, multiplier=1, fixedMin=5
+      // max(2*1, 5) = max(2, 5) = 5
+      expect(calculateMaxQueueSize(2, 1, 5)).toBe(5)
+    })
+
+    it('should handle zero multiplier', () => {
+      // concurrencyLimit=10, multiplier=0, fixedMin=3
+      // max(10*0, 3) = max(0, 3) = 3
+      expect(calculateMaxQueueSize(10, 0, 3)).toBe(3)
+    })
+
+    it('should handle fractional multiplier', () => {
+      // concurrencyLimit=10, multiplier=1.5, fixedMin=5
+      // max(10*1.5, 5) = max(15, 5) = 15
+      expect(calculateMaxQueueSize(10, 1.5, 5)).toBe(15)
+    })
+  })
+
+  describe('TTL Calculation', () => {
+    /**
+     * 排队计数器 TTL 计算（与 redis.js 中的实现一致）
+     */
+    function calculateQueueTtl(timeoutMs, bufferSeconds = 30) {
+      return Math.ceil(timeoutMs / 1000) + bufferSeconds
+    }
+
+    it('should calculate TTL with default buffer', () => {
+      // 60000ms = 60s + 30s buffer = 90s
+      expect(calculateQueueTtl(60000)).toBe(90)
+    })
+
+    it('should round up milliseconds to seconds', () => {
+      // 61500ms = ceil(61.5) = 62s + 30s = 92s
+      expect(calculateQueueTtl(61500)).toBe(92)
+    })
+
+    it('should handle custom buffer', () => {
+      // 30000ms = 30s + 60s buffer = 90s
+      expect(calculateQueueTtl(30000, 60)).toBe(90)
+    })
+
+    it('should handle very short timeout', () => {
+      // 1000ms = 1s + 30s = 31s
+      expect(calculateQueueTtl(1000)).toBe(31)
+    })
+  })
+})