fix: update totalTokens after compaction using last-call usage (#15018)

Merged via /review-pr -> /prepare-pr -> /merge-pr.

Prepared head SHA: 9214291bf7
Co-authored-by: shtse8 <8020099+shtse8@users.noreply.github.com>
Co-authored-by: gumadeiras <5599352+gumadeiras@users.noreply.github.com>
Reviewed-by: @gumadeiras

src/agents/pi-embedded-runner/run.ts

@@ -820,11 +820,18 @@ export async function runEmbeddedPiAgent(
           }
 
           const usage = toNormalizedUsage(usageAccumulator);
+          // Extract the last individual API call's usage for context-window
+          // utilization display. The accumulated `usage` sums input tokens
+          // across all calls (tool-use loops, compaction retries), which
+          // overstates the actual context size. `lastCallUsage` reflects only
+          // the final call, giving an accurate snapshot of current context.
+          const lastCallUsage = normalizeUsage(lastAssistant?.usage as UsageLike);
           const agentMeta: EmbeddedPiAgentMeta = {
             sessionId: sessionIdUsed,
             provider: lastAssistant?.provider ?? provider,
             model: lastAssistant?.model ?? model.id,
             usage,
+            lastCallUsage: lastCallUsage ?? undefined,
             compactionCount: autoCompactionCount > 0 ? autoCompactionCount : undefined,
           };
 

src/agents/pi-embedded-runner/types.ts

@@ -13,6 +13,20 @@ export type EmbeddedPiAgentMeta = {
     cacheWrite?: number;
     total?: number;
   };
+  /**
+   * Usage from the last individual API call (not accumulated across tool-use
+   * loops or compaction retries). Used for context-window utilization display
+   * (`totalTokens` in sessions.json) because the accumulated `usage.input`
+   * sums input tokens from every API call in the run, which overstates the
+   * actual context size.
+   */
+  lastCallUsage?: {
+    input?: number;
+    output?: number;
+    cacheRead?: number;
+    cacheWrite?: number;
+    total?: number;
+  };
 };
 
 export type EmbeddedPiRunMeta = {