fix(followup): prevent tool call UI leak and Enter accept buffer race

docs/design/prompt-suggestion/prompt-suggestion-design.md

@@ -64,10 +64,19 @@ A **prompt suggestion** (Next-step Suggestion / NES) is a short prediction (2-12
 ```
 [SUGGESTION MODE: Suggest what the user might naturally type next.]
 
+FIRST: Read the LAST FEW LINES of the assistant's most recent message — that's where
+next-step hints, tips, and actionable suggestions usually appear. Then check the user's
+recent messages and original request.
+
 Your job is to predict what THEY would type - not what you think they should do.
 THE TEST: Would they think "I was just about to type that"?
 
+PRIORITY: If the assistant's last message contains a tip or hint like "Tip: type X to ..."
+or "type X to ...", extract X as the suggestion. These are explicit next-step hints.
+
 EXAMPLES:
+Assistant says "Tip: type post comments to publish findings" → "post comments"
+Assistant says "type /review to start" → "/review"
 User asked "fix the bug and run tests", bug is fixed → "run the tests"
 After code written → "try it out"
 Task complete, obvious follow-up → "commit this" or "push it"
@@ -197,6 +206,23 @@ The Tab handler uses `key.name === 'tab'` explicitly (not `ACCEPT_SUGGESTION` ma
 | `enableSpeculation`         | boolean | false   | Predictive execution engine                                                      |
 | `fastModel` (top-level)     | string  | ""      | Model for all background tasks (empty = use main model). Set via `/model --fast` |
 
+### Internal Prompt ID Filtering
+
+Background operations use dedicated prompt IDs (`INTERNAL_PROMPT_IDS` in `utils/internalPromptIds.ts`) to prevent their API traffic and tool calls from appearing in the user-visible UI:
+
+| Prompt ID           | Used by                    |
+| ------------------- | -------------------------- |
+| `prompt_suggestion` | Suggestion generation      |
+| `forked_query`      | Cache-aware forked queries |
+| `speculation`       | Speculation engine         |
+
+**Filtering applied:**
+
+- `loggingContentGenerator` — skips `logApiRequest` and OpenAI interaction logging for internal IDs
+- `logApiResponse` / `logApiError` — skips `chatRecordingService.recordUiTelemetryEvent`
+- `logToolCall` — skips `chatRecordingService.recordUiTelemetryEvent`
+- `uiTelemetryService.addEvent` — **not filtered** (ensures `/stats` token tracking works)
+
 ### Thinking Mode
 
 Thinking/reasoning is explicitly disabled (`thinkingConfig: { includeThoughts: false }`) for all background task paths:

docs/users/features/followup-suggestions.md

@@ -12,7 +12,7 @@ After Qwen Code finishes responding, a suggestion appears as dimmed text in the
 > run the tests
 ```
 
-The suggestion is generated by sending the conversation history to the model, which predicts what you would naturally type next.
+The suggestion is generated by sending the conversation history to the model, which predicts what you would naturally type next. If the response contains an explicit tip (e.g., `Tip: type post comments to publish findings`), the suggested action is extracted automatically.
 
 ## Accepting Suggestions
 

packages/cli/src/ui/components/InputPrompt.test.tsx

@@ -235,6 +235,10 @@ describe('InputPrompt', () => {
       await wait();
 
       expect(props.onSubmit).toHaveBeenCalledWith('commit this');
+      // Enter path must NOT call buffer.insert — it passes text directly to
+      // handleSubmitAndClear. Calling insert would re-fill the buffer after
+      // it was already cleared (the microtask race bug).
+      expect(mockBuffer.insert).not.toHaveBeenCalled();
       unmount();
     });
 

packages/cli/src/ui/components/InputPrompt.tsx

@@ -892,7 +892,11 @@ export const InputPrompt: React.FC<InputPromptProps> = ({
           followup.state.suggestion
         ) {
           const text = followup.state.suggestion;
-          followup.accept('enter');
+          // Skip onAccept (buffer.insert) — we pass the text directly to
+          // handleSubmitAndClear which clears the buffer synchronously.
+          // Without skipOnAccept the microtask in accept() would re-insert
+          // the suggestion into the buffer after it was already cleared.
+          followup.accept('enter', { skipOnAccept: true });
           handleSubmitAndClear(text);
           return true;
         }

packages/cli/src/ui/hooks/useFollowupSuggestions.tsx

@@ -43,7 +43,10 @@ export interface UseFollowupSuggestionsReturn {
   /** Set suggestion text (called by parent component) */
   setSuggestion: (text: string | null) => void;
   /** Accept the current suggestion */
-  accept: (method?: 'tab' | 'enter' | 'right') => void;
+  accept: (
+    method?: 'tab' | 'enter' | 'right',
+    options?: { skipOnAccept?: boolean },
+  ) => void;
   /** Dismiss the current suggestion */
   dismiss: () => void;
   /** Clear all state */

packages/cli/src/ui/utils/MarkdownDisplay.tsx

@@ -111,7 +111,9 @@ const MarkdownDisplayInternal: React.FC<MarkdownDisplayProps> = ({
         lines[index + 1].match(tableSeparatorRegex)
       ) {
         inTable = true;
-        tableHeaders = tableRowMatch[1].split(/(?<!\\)\|/).map((cell) => cell.trim().replaceAll('\\|', '|'));
+        tableHeaders = tableRowMatch[1]
+          .split(/(?<!\\)\|/)
+          .map((cell) => cell.trim().replaceAll('\\|', '|'));
         tableRows = [];
       } else {
         // Not a table, treat as regular text
@@ -127,7 +129,9 @@ const MarkdownDisplayInternal: React.FC<MarkdownDisplayProps> = ({
       // Skip separator line - already handled
     } else if (inTable && tableRowMatch) {
       // Add table row
-      const cells = tableRowMatch[1].split(/(?<!\\)\|/).map((cell) => cell.trim().replaceAll('\\|', '|'));
+      const cells = tableRowMatch[1]
+        .split(/(?<!\\)\|/)
+        .map((cell) => cell.trim().replaceAll('\\|', '|'));
       // Ensure row has same column count as headers
       while (cells.length < tableHeaders.length) {
         cells.push('');

packages/cli/src/ui/utils/TableRenderer.tsx

@@ -36,7 +36,10 @@ export const TableRenderer: React.FC<TableRendererProps> = ({
   // Ensure table fits within terminal width
   const totalWidth = columnWidths.reduce((sum, width) => sum + width + 1, 1);
   const fixedWidth = columnWidths.length + 1;
-  const scaleFactor = totalWidth > contentWidth ? (contentWidth - fixedWidth) / (totalWidth - fixedWidth) : 1;
+  const scaleFactor =
+    totalWidth > contentWidth
+      ? (contentWidth - fixedWidth) / (totalWidth - fixedWidth)
+      : 1;
   const adjustedWidths = columnWidths.map((width) =>
     Math.floor(width * scaleFactor),
   );

packages/core/src/core/loggingContentGenerator/loggingContentGenerator.test.ts

@@ -23,11 +23,16 @@ import {
 import { OpenAILogger } from '../../utils/openaiLogger.js';
 import type OpenAI from 'openai';
 
-vi.mock('../../telemetry/loggers.js', () => ({
-  logApiRequest: vi.fn(),
-  logApiResponse: vi.fn(),
-  logApiError: vi.fn(),
-}));
+vi.mock('../../telemetry/loggers.js', async (importOriginal) => {
+  const actual =
+    await importOriginal<typeof import('../../telemetry/loggers.js')>();
+  return {
+    ...actual,
+    logApiRequest: vi.fn(),
+    logApiResponse: vi.fn(),
+    logApiError: vi.fn(),
+  };
+});
 
 vi.mock('../../utils/openaiLogger.js', () => ({
   OpenAILogger: vi.fn().mockImplementation(() => ({
@@ -474,4 +479,91 @@ describe('LoggingContentGenerator', () => {
       },
     ]);
   });
+
+  it.each(['prompt_suggestion', 'forked_query', 'speculation'])(
+    'skips logApiRequest and OpenAI logging for internal promptId %s (generateContent)',
+    async (promptId) => {
+      const mockResponse = {
+        responseId: 'internal-resp',
+        modelVersion: 'test-model',
+        candidates: [{ content: { parts: [{ text: 'suggestion' }] } }],
+        usageMetadata: { promptTokenCount: 10, candidatesTokenCount: 5 },
+      } as unknown as GenerateContentResponse;
+
+      const mockWrapped = {
+        generateContent: vi.fn().mockResolvedValue(mockResponse),
+        generateContentStream: vi.fn(),
+      } as unknown as ContentGenerator;
+
+      const gen = new LoggingContentGenerator(mockWrapped, createConfig(), {
+        model: 'test-model',
+        enableOpenAILogging: true,
+        openAILoggingDir: '/tmp/test-logs',
+      });
+
+      const request = {
+        model: 'test-model',
+        contents: [{ role: 'user', parts: [{ text: 'test' }] }],
+      } as unknown as GenerateContentParameters;
+
+      await gen.generateContent(request, promptId);
+
+      // logApiRequest should NOT be called for internal prompts
+      expect(logApiRequest).not.toHaveBeenCalled();
+      // logApiResponse SHOULD be called (for /stats token tracking)
+      expect(logApiResponse).toHaveBeenCalled();
+      // OpenAI logger should be constructed, but no interaction should be logged
+      expect(OpenAILogger).toHaveBeenCalled();
+      const loggerInstance = (
+        OpenAILogger as unknown as ReturnType<typeof vi.fn>
+      ).mock.results[0]?.value;
+      expect(loggerInstance.logInteraction).not.toHaveBeenCalled();
+    },
+  );
+
+  it.each(['prompt_suggestion', 'forked_query', 'speculation'])(
+    'skips logApiRequest and OpenAI logging for internal promptId %s (generateContentStream)',
+    async (promptId) => {
+      const mockChunk = {
+        responseId: 'stream-resp',
+        modelVersion: 'test-model',
+        candidates: [{ content: { parts: [{ text: 'suggestion' }] } }],
+        usageMetadata: { promptTokenCount: 10, candidatesTokenCount: 5 },
+      } as unknown as GenerateContentResponse;
+
+      async function* fakeStream() {
+        yield mockChunk;
+      }
+
+      const mockWrapped = {
+        generateContent: vi.fn(),
+        generateContentStream: vi.fn().mockResolvedValue(fakeStream()),
+      } as unknown as ContentGenerator;
+
+      const gen = new LoggingContentGenerator(mockWrapped, createConfig(), {
+        model: 'test-model',
+        enableOpenAILogging: true,
+        openAILoggingDir: '/tmp/test-logs',
+      });
+
+      const request = {
+        model: 'test-model',
+        contents: [{ role: 'user', parts: [{ text: 'test' }] }],
+      } as unknown as GenerateContentParameters;
+
+      const stream = await gen.generateContentStream(request, promptId);
+      // Consume the stream
+      for await (const _chunk of stream) {
+        // drain
+      }
+
+      expect(logApiRequest).not.toHaveBeenCalled();
+      expect(logApiResponse).toHaveBeenCalled();
+      expect(OpenAILogger).toHaveBeenCalled();
+      const loggerInstance = (
+        OpenAILogger as unknown as ReturnType<typeof vi.fn>
+      ).mock.results[0]?.value;
+      expect(loggerInstance.logInteraction).not.toHaveBeenCalled();
+    },
+  );
 });

packages/core/src/core/loggingContentGenerator/loggingContentGenerator.ts

@@ -31,6 +31,7 @@ import {
   logApiRequest,
   logApiResponse,
 } from '../../telemetry/loggers.js';
+import { isInternalPromptId } from '../../utils/internalPromptIds.js';
 import type {
   ContentGenerator,
   ContentGeneratorConfig,
@@ -143,8 +144,17 @@ export class LoggingContentGenerator implements ContentGenerator {
     userPromptId: string,
   ): Promise<GenerateContentResponse> {
     const startTime = Date.now();
-    this.logApiRequest(this.toContents(req.contents), req.model, userPromptId);
-    const openaiRequest = await this.buildOpenAIRequestForLogging(req);
+    const isInternal = isInternalPromptId(userPromptId);
+    if (!isInternal) {
+      this.logApiRequest(
+        this.toContents(req.contents),
+        req.model,
+        userPromptId,
+      );
+    }
+    const openaiRequest = isInternal
+      ? undefined
+      : await this.buildOpenAIRequestForLogging(req);
     try {
       const response = await this.wrapped.generateContent(req, userPromptId);
       const durationMs = Date.now() - startTime;
@@ -155,12 +165,16 @@ export class LoggingContentGenerator implements ContentGenerator {
         userPromptId,
         response.usageMetadata,
       );
-      await this.logOpenAIInteraction(openaiRequest, response);
+      if (!isInternal) {
+        await this.logOpenAIInteraction(openaiRequest, response);
+      }
       return response;
     } catch (error) {
       const durationMs = Date.now() - startTime;
       this._logApiError('', durationMs, error, req.model, userPromptId);
-      await this.logOpenAIInteraction(openaiRequest, undefined, error);
+      if (!isInternal) {
+        await this.logOpenAIInteraction(openaiRequest, undefined, error);
+      }
       throw error;
     }
   }
@@ -170,16 +184,27 @@ export class LoggingContentGenerator implements ContentGenerator {
     userPromptId: string,
   ): Promise<AsyncGenerator<GenerateContentResponse>> {
     const startTime = Date.now();
-    this.logApiRequest(this.toContents(req.contents), req.model, userPromptId);
-    const openaiRequest = await this.buildOpenAIRequestForLogging(req);
+    const isInternal = isInternalPromptId(userPromptId);
+    if (!isInternal) {
+      this.logApiRequest(
+        this.toContents(req.contents),
+        req.model,
+        userPromptId,
+      );
+    }
+    const openaiRequest = isInternal
+      ? undefined
+      : await this.buildOpenAIRequestForLogging(req);
 
     let stream: AsyncGenerator<GenerateContentResponse>;
     try {
       stream = await this.wrapped.generateContentStream(req, userPromptId);
     } catch (error) {
       const durationMs = Date.now() - startTime;
       this._logApiError('', durationMs, error, req.model, userPromptId);
-      await this.logOpenAIInteraction(openaiRequest, undefined, error);
+      if (!isInternal) {
+        await this.logOpenAIInteraction(openaiRequest, undefined, error);
+      }
       throw error;
     }
 
@@ -199,12 +224,27 @@ export class LoggingContentGenerator implements ContentGenerator {
     model: string,
     openaiRequest?: OpenAI.Chat.ChatCompletionCreateParams,
   ): AsyncGenerator<GenerateContentResponse> {
+    const isInternal = isInternalPromptId(userPromptId);
+    // For internal prompts we only need the last usage metadata (for /stats);
+    // skip collecting full responses to avoid unnecessary memory overhead.
     const responses: GenerateContentResponse[] = [];
 
+    // Track first-seen IDs so _logApiResponse/_logApiError have accurate
+    // values even when we skip collecting full responses for internal prompts.
+    let firstResponseId = '';
+    let firstModelVersion = '';
     let lastUsageMetadata: GenerateContentResponseUsageMetadata | undefined;
     try {
       for await (const response of stream) {
-        responses.push(response);
+        if (!firstResponseId && response.responseId) {
+          firstResponseId = response.responseId;
+        }
+        if (!firstModelVersion && response.modelVersion) {
+          firstModelVersion = response.modelVersion;
+        }
+        if (!isInternal) {
+          responses.push(response);
+        }
         if (response.usageMetadata) {
           lastUsageMetadata = response.usageMetadata;
         }
@@ -213,25 +253,29 @@ export class LoggingContentGenerator implements ContentGenerator {
       // Only log successful API response if no error occurred
       const durationMs = Date.now() - startTime;
       this._logApiResponse(
-        responses[0]?.responseId ?? '',
+        firstResponseId,
         durationMs,
-        responses[0]?.modelVersion || model,
+        firstModelVersion || model,
         userPromptId,
         lastUsageMetadata,
       );
-      const consolidatedResponse =
-        this.consolidateGeminiResponsesForLogging(responses);
-      await this.logOpenAIInteraction(openaiRequest, consolidatedResponse);
+      if (!isInternal) {
+        const consolidatedResponse =
+          this.consolidateGeminiResponsesForLogging(responses);
+        await this.logOpenAIInteraction(openaiRequest, consolidatedResponse);
+      }
     } catch (error) {
       const durationMs = Date.now() - startTime;
       this._logApiError(
-        responses[0]?.responseId ?? '',
+        firstResponseId,
         durationMs,
         error,
-        responses[0]?.modelVersion || model,
+        firstModelVersion || model,
         userPromptId,
       );
-      await this.logOpenAIInteraction(openaiRequest, undefined, error);
+      if (!isInternal) {
+        await this.logOpenAIInteraction(openaiRequest, undefined, error);
+      }
       throw error;
     }
   }