diff --git a/src/agents/session-tool-result-guard.test.ts b/src/agents/session-tool-result-guard.test.ts
index 7b656606646..105b587cf9d 100644
--- a/src/agents/session-tool-result-guard.test.ts
+++ b/src/agents/session-tool-result-guard.test.ts
@@ -357,4 +357,63 @@ describe("installSessionToolResultGuard", () => {
       sourceTool: "sessions_send",
     });
   });
+
+  // Regression test for orphaned tool_result bug
+  // See: https://github.com/clawdbot/clawdbot/issues/XXXX
+  // When an assistant message with toolCalls is aborted, no synthetic toolResult
+  // should be created. Creating synthetic results for aborted/incomplete tool calls
+  // causes API 400 errors: "unexpected tool_use_id found in tool_result blocks"
+  it("does NOT create synthetic toolResult for aborted assistant messages with toolCalls", () => {
+    const sm = SessionManager.inMemory();
+    installSessionToolResultGuard(sm);
+
+    // Aborted assistant message with incomplete toolCall
+    sm.appendMessage(
+      asAppendMessage({
+        role: "assistant",
+        content: [{ type: "toolCall", id: "call_aborted", name: "read", arguments: {} }],
+        stopReason: "aborted",
+      }),
+    );
+
+    // Next message triggers flush of pending tool calls
+    sm.appendMessage(
+      asAppendMessage({
+        role: "user",
+        content: "are you stuck?",
+        timestamp: Date.now(),
+      }),
+    );
+
+    // Should only have assistant + user, NO synthetic toolResult
+    const messages = getPersistedMessages(sm);
+    const roles = messages.map((m) => m.role);
+    expect(roles).toEqual(["assistant", "user"]);
+    expect(roles).not.toContain("toolResult");
+  });
+
+  it("does NOT create synthetic toolResult for errored assistant messages with toolCalls", () => {
+    const sm = SessionManager.inMemory();
+    const guard = installSessionToolResultGuard(sm);
+
+    // Error assistant message with incomplete toolCall
+    sm.appendMessage(
+      asAppendMessage({
+        role: "assistant",
+        content: [{ type: "toolCall", id: "call_error", name: "exec", arguments: {} }],
+        stopReason: "error",
+      }),
+    );
+
+    // Explicit flush should NOT create synthetic result for errored messages
+    guard.flushPendingToolResults();
+
+    const messages = getPersistedMessages(sm);
+    const toolResults = messages.filter((m) => m.role === "toolResult");
+    // No synthetic toolResults should exist for the errored call
+    const syntheticForError = toolResults.filter(
+      (m) => (m as { toolCallId?: string }).toolCallId === "call_error",
+    );
+    expect(syntheticForError).toHaveLength(0);
+  });
 });
diff --git a/src/agents/session-tool-result-guard.ts b/src/agents/session-tool-result-guard.ts
index 689bb816c1e..dba618a3103 100644
--- a/src/agents/session-tool-result-guard.ts
+++ b/src/agents/session-tool-result-guard.ts
@@ -166,8 +166,15 @@ export function installSessionToolResultGuard(
       return originalAppend(persisted as never);
     }
 
+    // Skip tool call extraction for aborted/errored assistant messages.
+    // When stopReason is "error" or "aborted", the tool_use blocks may be incomplete
+    // and should not have synthetic tool_results created. Creating synthetic results
+    // for incomplete tool calls causes API 400 errors:
+    // "unexpected tool_use_id found in tool_result blocks"
+    // This matches the behavior in repairToolUseResultPairing (session-transcript-repair.ts)
+    const stopReason = (nextMessage as { stopReason?: string }).stopReason;
     const toolCalls =
-      nextRole === "assistant"
+      nextRole === "assistant" && stopReason !== "aborted" && stopReason !== "error"
         ? extractToolCallsFromAssistant(nextMessage as Extract<AgentMessage, { role: "assistant" }>)
         : [];