feat: update tool output format, use plain string instead of json string (#881)

packages/cli/src/ui/AppContainer.tsx

@@ -555,7 +555,7 @@ export const AppContainer = (props: AppContainerProps) => {
     historyManager.addItem(
       {
         type: MessageType.INFO,
-        text: 'Refreshing hierarchical memory (GEMINI.md or other context files)...',
+        text: 'Refreshing hierarchical memory (QWEN.md or other context files)...',
       },
       Date.now(),
     );

packages/core/src/core/openaiContentGenerator/converter.test.ts

@@ -7,6 +7,7 @@
 import { describe, it, expect, beforeEach } from 'vitest';
 import { OpenAIContentConverter } from './converter.js';
 import type { StreamingToolCallParser } from './streamingToolCallParser.js';
+import type { GenerateContentParameters, Content } from '@google/genai';
 
 describe('OpenAIContentConverter', () => {
   let converter: OpenAIContentConverter;
@@ -68,4 +69,77 @@ describe('OpenAIContentConverter', () => {
       expect(parser.getBuffer(0)).toBe('');
     });
   });
+
+  describe('convertGeminiRequestToOpenAI', () => {
+    const createRequestWithFunctionResponse = (
+      response: Record<string, unknown>,
+    ): GenerateContentParameters => {
+      const contents: Content[] = [
+        {
+          role: 'model',
+          parts: [
+            {
+              functionCall: {
+                id: 'call_1',
+                name: 'shell',
+                args: {},
+              },
+            },
+          ],
+        },
+        {
+          role: 'user',
+          parts: [
+            {
+              functionResponse: {
+                id: 'call_1',
+                name: 'shell',
+                response,
+              },
+            },
+          ],
+        },
+      ];
+      return {
+        model: 'models/test',
+        contents,
+      };
+    };
+
+    it('should extract raw output from function response objects', () => {
+      const request = createRequestWithFunctionResponse({
+        output: 'Raw output text',
+      });
+
+      const messages = converter.convertGeminiRequestToOpenAI(request);
+      const toolMessage = messages.find((message) => message.role === 'tool');
+
+      expect(toolMessage).toBeDefined();
+      expect(toolMessage?.content).toBe('Raw output text');
+    });
+
+    it('should prioritize error field when present', () => {
+      const request = createRequestWithFunctionResponse({
+        error: 'Command failed',
+      });
+
+      const messages = converter.convertGeminiRequestToOpenAI(request);
+      const toolMessage = messages.find((message) => message.role === 'tool');
+
+      expect(toolMessage).toBeDefined();
+      expect(toolMessage?.content).toBe('Command failed');
+    });
+
+    it('should stringify non-string responses', () => {
+      const request = createRequestWithFunctionResponse({
+        data: { value: 42 },
+      });
+
+      const messages = converter.convertGeminiRequestToOpenAI(request);
+      const toolMessage = messages.find((message) => message.role === 'tool');
+
+      expect(toolMessage).toBeDefined();
+      expect(toolMessage?.content).toBe('{"data":{"value":42}}');
+    });
+  });
 });

packages/core/src/core/openaiContentGenerator/converter.ts

@@ -276,10 +276,7 @@ export class OpenAIContentConverter {
         messages.push({
           role: 'tool' as const,
           tool_call_id: funcResponse.id || '',
-          content:
+          content: this.extractFunctionResponseContent(funcResponse.response),
-            typeof funcResponse.response === 'string'
-              ? funcResponse.response
-              : JSON.stringify(funcResponse.response),
         });
       }
       return;
@@ -359,6 +356,36 @@ export class OpenAIContentConverter {
     return { textParts, functionCalls, functionResponses, mediaParts };
   }
 
+  private extractFunctionResponseContent(response: unknown): string {
+    if (response === null || response === undefined) {
+      return '';
+    }
+
+    if (typeof response === 'string') {
+      return response;
+    }
+
+    if (typeof response === 'object') {
+      const responseObject = response as Record<string, unknown>;
+      const output = responseObject['output'];
+      if (typeof output === 'string') {
+        return output;
+      }
+
+      const error = responseObject['error'];
+      if (typeof error === 'string') {
+        return error;
+      }
+    }
+
+    try {
+      const serialized = JSON.stringify(response);
+      return serialized ?? String(response);
+    } catch {
+      return String(response);
+    }
+  }
+
   /**
    * Determine media type from MIME type
    */

packages/core/src/tools/exitPlanMode.test.ts

@@ -131,16 +131,14 @@ describe('ExitPlanModeTool', () => {
       }
 
       const result = await invocation.execute(signal);
-      const expectedLlmMessage =
-        'User has approved your plan. You can now start coding. Start with updating your todo list if applicable.';
 
-      expect(result).toEqual({
+      expect(result.llmContent).toContain(
-        llmContent: expectedLlmMessage,
+        'User has approved your plan. You can now start coding',
-        returnDisplay: {
+      );
-          type: 'plan_summary',
+      expect(result.returnDisplay).toEqual({
-          message: 'User approved the plan.',
+        type: 'plan_summary',
-          plan: params.plan,
+        message: 'User approved the plan.',
-        },
+        plan: params.plan,
       });
 
       expect(mockConfig.setApprovalMode).toHaveBeenCalledWith(
@@ -188,15 +186,12 @@ describe('ExitPlanModeTool', () => {
 
       const result = await invocation.execute(signal);
 
-      expect(result).toEqual({
+      expect(result.llmContent).toBe(
-        llmContent: JSON.stringify({
+        'Plan execution was not approved. Remaining in plan mode.',
-          success: false,
+      );
-          plan: params.plan,
+      expect(result.returnDisplay).toBe(
-          error: 'Plan execution was not approved. Remaining in plan mode.',
+        'Plan execution was not approved. Remaining in plan mode.',
-        }),
+      );
-        returnDisplay:
-          'Plan execution was not approved. Remaining in plan mode.',
-      });
 
       expect(mockConfig.setApprovalMode).toHaveBeenCalledWith(
         ApprovalMode.PLAN,
@@ -215,50 +210,6 @@ describe('ExitPlanModeTool', () => {
       );
     });
 
-    it('should handle execution errors gracefully', async () => {
-      const params: ExitPlanModeParams = {
-        plan: 'Test plan',
-      };
-
-      const invocation = tool.build(params);
-      const confirmation = await invocation.shouldConfirmExecute(
-        new AbortController().signal,
-      );
-      if (confirmation) {
-        // Don't approve the plan so we go through the rejection path
-        await confirmation.onConfirm(ToolConfirmationOutcome.Cancel);
-      }
-
-      // Create a spy to simulate an error during the execution
-      const consoleSpy = vi
-        .spyOn(console, 'error')
-        .mockImplementation(() => {});
-
-      // Mock JSON.stringify to throw an error in the rejection path
-      const originalStringify = JSON.stringify;
-      vi.spyOn(JSON, 'stringify').mockImplementationOnce(() => {
-        throw new Error('JSON stringify error');
-      });
-
-      const result = await invocation.execute(new AbortController().signal);
-
-      expect(result).toEqual({
-        llmContent: JSON.stringify({
-          success: false,
-          error: 'Failed to present plan. Detail: JSON stringify error',
-        }),
-        returnDisplay: 'Error presenting plan: JSON stringify error',
-      });
-
-      expect(consoleSpy).toHaveBeenCalledWith(
-        '[ExitPlanModeTool] Error executing exit_plan_mode: JSON stringify error',
-      );
-
-      // Restore original JSON.stringify
-      JSON.stringify = originalStringify;
-      consoleSpy.mockRestore();
-    });
-
     it('should return empty tool locations', () => {
       const params: ExitPlanModeParams = {
         plan: 'Test plan',

packages/core/src/tools/exitPlanMode.ts

@@ -115,17 +115,12 @@ class ExitPlanModeToolInvocation extends BaseToolInvocation<
         const rejectionMessage =
           'Plan execution was not approved. Remaining in plan mode.';
         return {
-          llmContent: JSON.stringify({
+          llmContent: rejectionMessage,
-            success: false,
-            plan,
-            error: rejectionMessage,
-          }),
           returnDisplay: rejectionMessage,
         };
       }
 
-      const llmMessage =
+      const llmMessage = `User has approved your plan. You can now start coding. Start with updating your todo list if applicable.`;
-        'User has approved your plan. You can now start coding. Start with updating your todo list if applicable.';
       const displayMessage = 'User approved the plan.';
 
       return {
@@ -142,11 +137,11 @@ class ExitPlanModeToolInvocation extends BaseToolInvocation<
       console.error(
         `[ExitPlanModeTool] Error executing exit_plan_mode: ${errorMessage}`,
       );
+
+      const errorLlmContent = `Failed to present plan: ${errorMessage}`;
+
       return {
-        llmContent: JSON.stringify({
+        llmContent: errorLlmContent,
-          success: false,
-          error: `Failed to present plan. Detail: ${errorMessage}`,
-        }),
         returnDisplay: `Error presenting plan: ${errorMessage}`,
       };
     }

packages/core/src/tools/memoryTool.test.ts

@@ -241,9 +241,7 @@ describe('MemoryTool', () => {
         expectedFsArgument,
       );
       const successMessage = `Okay, I've remembered that in global memory: "${params.fact}"`;
-      expect(result.llmContent).toBe(
+      expect(result.llmContent).toBe(successMessage);
-        JSON.stringify({ success: true, message: successMessage }),
-      );
       expect(result.returnDisplay).toBe(successMessage);
     });
 
@@ -271,9 +269,7 @@ describe('MemoryTool', () => {
         expectedFsArgument,
       );
       const successMessage = `Okay, I've remembered that in project memory: "${params.fact}"`;
-      expect(result.llmContent).toBe(
+      expect(result.llmContent).toBe(successMessage);
-        JSON.stringify({ success: true, message: successMessage }),
-      );
       expect(result.returnDisplay).toBe(successMessage);
     });
 
@@ -298,10 +294,7 @@ describe('MemoryTool', () => {
       const result = await invocation.execute(mockAbortSignal);
 
       expect(result.llmContent).toBe(
-        JSON.stringify({
+        `Error saving memory: ${underlyingError.message}`,
-          success: false,
-          error: `Failed to save memory. Detail: ${underlyingError.message}`,
-        }),
       );
       expect(result.returnDisplay).toBe(
         `Error saving memory: ${underlyingError.message}`,
@@ -319,6 +312,8 @@ describe('MemoryTool', () => {
       expect(result.llmContent).toContain(
         'Please specify where to save this memory',
       );
+      expect(result.llmContent).toContain('Global:');
+      expect(result.llmContent).toContain('Project:');
       expect(result.returnDisplay).toContain('Global:');
       expect(result.returnDisplay).toContain('Project:');
     });

packages/core/src/tools/memoryTool.ts

@@ -309,7 +309,7 @@ Preview of changes to be made to GLOBAL memory:
     if (!fact || typeof fact !== 'string' || fact.trim() === '') {
       const errorMessage = 'Parameter "fact" must be a non-empty string.';
       return {
-        llmContent: JSON.stringify({ success: false, error: errorMessage }),
+        llmContent: `Error: ${errorMessage}`,
         returnDisplay: `Error: ${errorMessage}`,
       };
     }
@@ -324,10 +324,7 @@ Global: ${globalPath} (shared across all projects)
 Project: ${projectPath} (current project only)`;
 
       return {
-        llmContent: JSON.stringify({
+        llmContent: errorMessage,
-          success: false,
-          error: 'Please specify where to save this memory',
-        }),
         returnDisplay: errorMessage,
       };
     }
@@ -344,10 +341,7 @@ Project: ${projectPath} (current project only)`;
         await fs.writeFile(memoryFilePath, modified_content, 'utf-8');
         const successMessage = `Okay, I've updated the ${scope} memory file with your modifications.`;
         return {
-          llmContent: JSON.stringify({
+          llmContent: successMessage,
-            success: true,
-            message: successMessage,
-          }),
           returnDisplay: successMessage,
         };
       } else {
@@ -359,10 +353,7 @@ Project: ${projectPath} (current project only)`;
         });
         const successMessage = `Okay, I've remembered that in ${scope} memory: "${fact}"`;
         return {
-          llmContent: JSON.stringify({
+          llmContent: successMessage,
-            success: true,
-            message: successMessage,
-          }),
           returnDisplay: successMessage,
         };
       }
@@ -372,11 +363,9 @@ Project: ${projectPath} (current project only)`;
       console.error(
         `[MemoryTool] Error executing save_memory for fact "${fact}" in ${scope}: ${errorMessage}`,
       );
+
       return {
-        llmContent: JSON.stringify({
+        llmContent: `Error saving memory: ${errorMessage}`,
-          success: false,
-          error: `Failed to save memory. Detail: ${errorMessage}`,
-        }),
         returnDisplay: `Error saving memory: ${errorMessage}`,
         error: {
           message: errorMessage,

packages/core/src/tools/todoWrite.test.ts

@@ -141,7 +141,12 @@ describe('TodoWriteTool', () => {
       const invocation = tool.build(params);
       const result = await invocation.execute(mockAbortSignal);
 
-      expect(result.llmContent).toContain('success');
+      expect(result.llmContent).toContain(
+        'Todos have been modified successfully',
+      );
+      expect(result.llmContent).toContain('<system-reminder>');
+      expect(result.llmContent).toContain('Your todo list has changed');
+      expect(result.llmContent).toContain(JSON.stringify(params.todos));
       expect(result.returnDisplay).toEqual({
         type: 'todo_list',
         todos: [
@@ -178,7 +183,12 @@ describe('TodoWriteTool', () => {
       const invocation = tool.build(params);
       const result = await invocation.execute(mockAbortSignal);
 
-      expect(result.llmContent).toContain('success');
+      expect(result.llmContent).toContain(
+        'Todos have been modified successfully',
+      );
+      expect(result.llmContent).toContain('<system-reminder>');
+      expect(result.llmContent).toContain('Your todo list has changed');
+      expect(result.llmContent).toContain(JSON.stringify(params.todos));
       expect(result.returnDisplay).toEqual({
         type: 'todo_list',
         todos: [
@@ -208,7 +218,10 @@ describe('TodoWriteTool', () => {
       const invocation = tool.build(params);
       const result = await invocation.execute(mockAbortSignal);
 
-      expect(result.llmContent).toContain('"success":false');
+      expect(result.llmContent).toContain('Failed to modify todos');
+      expect(result.llmContent).toContain('<system-reminder>');
+      expect(result.llmContent).toContain('Todo list modification failed');
+      expect(result.llmContent).toContain('Write failed');
       expect(result.returnDisplay).toContain('Error writing todos');
     });
 
@@ -223,7 +236,10 @@ describe('TodoWriteTool', () => {
       const invocation = tool.build(params);
       const result = await invocation.execute(mockAbortSignal);
 
-      expect(result.llmContent).toContain('success');
+      expect(result.llmContent).toContain('Todo list has been cleared');
+      expect(result.llmContent).toContain('<system-reminder>');
+      expect(result.llmContent).toContain('Your todo list is now empty');
+      expect(result.llmContent).toContain('no pending tasks');
       expect(result.returnDisplay).toEqual({
         type: 'todo_list',
         todos: [],

packages/core/src/tools/todoWrite.ts

@@ -340,11 +340,30 @@ class TodoWriteToolInvocation extends BaseToolInvocation<
         todos: finalTodos,
       };
 
+      // Create plain string format with system reminder
+      const todosJson = JSON.stringify(finalTodos);
+      let llmContent: string;
+
+      if (finalTodos.length === 0) {
+        // Special message for empty todos
+        llmContent = `Todo list has been cleared.
+
+<system-reminder>
+Your todo list is now empty. DO NOT mention this explicitly to the user. You have no pending tasks in your todo list.
+</system-reminder>`;
+      } else {
+        // Normal message for todos with items
+        llmContent = `Todos have been modified successfully. Ensure that you continue to use the todo list to track your progress. Please proceed with the current tasks if applicable
+
+<system-reminder>
+Your todo list has changed. DO NOT mention this explicitly to the user. Here are the latest contents of your todo list: 
+
+${todosJson}. Continue on with the tasks at hand if applicable.
+</system-reminder>`;
+      }
+
       return {
-        llmContent: JSON.stringify({
+        llmContent,
-          success: true,
-          todos: finalTodos,
-        }),
         returnDisplay: todoResultDisplay,
       };
     } catch (error) {
@@ -353,11 +372,16 @@ class TodoWriteToolInvocation extends BaseToolInvocation<
       console.error(
         `[TodoWriteTool] Error executing todo_write: ${errorMessage}`,
       );
+
+      // Create plain string format for error with system reminder
+      const errorLlmContent = `Failed to modify todos. An error occurred during the operation.
+
+<system-reminder>
+Todo list modification failed with error: ${errorMessage}. You may need to retry or handle this error appropriately.
+</system-reminder>`;
+
       return {
-        llmContent: JSON.stringify({
+        llmContent: errorLlmContent,
-          success: false,
-          error: `Failed to write todos. Detail: ${errorMessage}`,
-        }),
         returnDisplay: `Error writing todos: ${errorMessage}`,
       };
     }