fix: prevent subagent telemetry from overwriting main agent footer context

packages/core/src/core/geminiChat.test.ts

@@ -788,6 +788,48 @@ describe('GeminiChat', async () => {
       );
     });
 
+    it('should not update global telemetry when no telemetryService is provided (subagent isolation)', async () => {
+      // Simulate a subagent GeminiChat: created without a telemetryService
+      const subagentChat = new GeminiChat(mockConfig, config, []);
+
+      const response = (async function* () {
+        yield {
+          candidates: [
+            {
+              content: {
+                parts: [{ text: 'subagent response' }],
+                role: 'model',
+              },
+              finishReason: 'STOP',
+              index: 0,
+              safetyRatings: [],
+            },
+          ],
+          text: () => 'subagent response',
+          usageMetadata: {
+            promptTokenCount: 12000,
+            candidatesTokenCount: 500,
+            totalTokenCount: 12500,
+          },
+        } as unknown as GenerateContentResponse;
+      })();
+      vi.mocked(mockContentGenerator.generateContentStream).mockResolvedValue(
+        response,
+      );
+
+      const stream = await subagentChat.sendMessageStream(
+        'test-model',
+        { message: 'subagent task' },
+        'prompt-id-subagent',
+      );
+      for await (const _ of stream) {
+        // consume stream
+      }
+
+      // The global uiTelemetryService must NOT be called by subagent chats
+      expect(uiTelemetryService.setLastPromptTokenCount).not.toHaveBeenCalled();
+    });
+
     it('should keep parts with thoughtSignature when consolidating history', async () => {
       const stream = (async function* () {
         yield {

packages/core/src/core/geminiChat.ts

@@ -35,7 +35,6 @@ import {
   ContentRetryFailureEvent,
 } from '../telemetry/types.js';
 import type { UiTelemetryService } from '../telemetry/uiTelemetry.js';
-import { uiTelemetryService } from '../telemetry/uiTelemetry.js';
 
 const debugLogger = createDebugLogger('QWEN_CODE_CHAT');
 
@@ -659,15 +658,11 @@ export class GeminiChat {
         // Some providers omit total_tokens or return 0 in streaming usage chunks.
         const lastPromptTokenCount =
           usageMetadata.totalTokenCount || usageMetadata.promptTokenCount;
-        if (lastPromptTokenCount) {
-          (this.telemetryService ?? uiTelemetryService).setLastPromptTokenCount(
-            lastPromptTokenCount,
-          );
+        if (lastPromptTokenCount && this.telemetryService) {
+          this.telemetryService.setLastPromptTokenCount(lastPromptTokenCount);
         }
-        if (usageMetadata.cachedContentTokenCount) {
-          (
-            this.telemetryService ?? uiTelemetryService
-          ).setLastCachedContentTokenCount(
+        if (usageMetadata.cachedContentTokenCount && this.telemetryService) {
+          this.telemetryService.setLastCachedContentTokenCount(
             usageMetadata.cachedContentTokenCount,
           );
         }