AI Conversation Logger

Instructions

Implementation Reference

• src/tools/conversationLogger.ts:25-115 (handler)

  The logConversation method implements the core tool logic: validates parameters, auto-detects project, generates conversation entry, appends to daily Markdown log file, and returns formatted success or error response.

  async logConversation(params: unknown): Promise<{ content: Array<{ type: string; text: string }> }> {
    try {
      const validatedParams = validateLogConversation(params);
      const {
        userRequest,
        aiTodoList,
        aiSummary,
        fileOperations,
        title,
        tags,
        platform
      } = validatedParams;
  
      // Auto-detect project from current working directory
      const projectInfo = await this.fileManager.getProjectInfo();
      const projectName = projectInfo.name;
  
      await this.fileManager.initializeProject();
  
      const now = new Date();
      // Use zero-padded date format for better sorting (2025-08-07 instead of 2025-8-7)
      const year = now.getFullYear();
      const month = String(now.getMonth() + 1).padStart(2, '0');
      const day = String(now.getDate()).padStart(2, '0');
      const dateStr = `${year}-${month}-${day}`;
      const conversationEntry: ConversationEntry = {
        id: randomUUID(),
        timestamp: now.toISOString(),
        project: projectName,
        platform: platform || CONSTANTS.PLATFORMS.UNKNOWN,
        userRequest: userRequest.trim(),
        aiTodoList: aiTodoList || [],
        aiSummary: aiSummary.trim(),
        fileOperations: fileOperations || [],
        title: title?.trim() || undefined,
        tags: tags || []
      };
  
      const dailyLogPath = await this.fileManager.getDailyLogPath(projectName, dateStr);
      let existingContent = await this.fileManager.readFile(dailyLogPath);
      
      // Calculate conversation number
      const existingConversationCount = (existingContent.match(/## 对话/g) || []).length;
      const conversationNumber = existingConversationCount + 1;
      
      // Generate the conversation markdown
      const conversationMarkdown = MarkdownFormatter.formatConversationEntry(conversationEntry, conversationNumber);
      
      if (!existingContent) {
        // New file: create header + new conversation
        const header = MarkdownFormatter.formatDailyLogHeader(projectName, dateStr, projectInfo.root);
        const finalContent = header + conversationMarkdown;
        await this.fileManager.writeFile(dailyLogPath, finalContent);
      } else {
        // Existing file: just append new conversation
        const finalContent = existingContent + conversationMarkdown;
        await this.fileManager.writeFile(dailyLogPath, finalContent);
      }
  
      return {
        content: [
          {
            type: 'text',
            text: `✅ 对话已成功记录到项目 "${projectName}"！\n\n` +
                  `📁 文件位置: ${dailyLogPath}\n` +
                  `💻 项目目录: ${projectInfo.root || '未检测到项目'}\n` +
                  `🕒 时间: ${now.toLocaleString('zh-CN')}\n` +
                  `🏷️ 标签: ${(tags && tags.length > 0) ? tags.join(', ') : '无'}\n` +
                  `⚡ 执行任务: ${aiTodoList?.length || 0}\n` +
                  `📂 文件操作: ${fileOperations?.length || 0}\n` +
                  `📝 对话编号: 对话${conversationNumber}`
          }
        ]
      };
    } catch (error) {
      const errorMessage = error instanceof ValidationError 
        ? `参数验证失败: ${error.message}`
        : error instanceof FileOperationError
        ? `文件操作失败: ${error.message} (文件: ${error.filePath})`
        : `记录对话时出错: ${String(error)}`;
      
      return {
        content: [
          {
            type: 'text',
            text: `❌ ${errorMessage}`
          }
        ]
      };
    }
  }

• src/validation/schemas.ts:19-33 (schema)

  Zod schema for validating input parameters of the log_conversation tool, including transformation for defaults.

  export const LogConversationSchema = z.object({
    project: z.string().optional(),  // 项目名称(可选,自动检测)
    userRequest: z.string().min(1, '用户需求不能为空'),
    aiTodoList: z.array(z.string()).min(1, 'AI执行计划不能为空'),
    aiSummary: z.string().min(1, 'AI总结不能为空'),
    fileOperations: z.array(z.string()).optional(),
    title: z.string().optional(),
    tags: z.array(z.string()).optional(),
    platform: z.string().default(CONSTANTS.PLATFORMS.CLAUDE_CODE)
  }).transform((data) => ({
    ...data,
    tags: data.tags || [],
    fileOperations: data.fileOperations || [],
    platform: data.platform || CONSTANTS.PLATFORMS.CLAUDE_CODE
  }));

• src/index.ts:38-78 (registration)

  MCP tool registration: defines the 'log_conversation' tool specification including name, description, and input schema in the ListTools response.

  {
    name: 'log_conversation',
    description: '记录AI对话 - 所有会话都要记录\n\n使用规范：\n• userRequest: 用户原始需求+上传文件说明\n• aiTodoList: 你的执行计划清单(即使只是查看也要列出)\n• aiSummary: 你的操作总结(3-5句话,包括解释、分析等)\n• fileOperations: 文件操作总结,格式:"动作 文件路径 - 说明"(可为空)\n• title: 对话标题(可选)\n• tags: 标签数组(可选)',
    inputSchema: {
      type: 'object',
      properties: {
        userRequest: {
          type: 'string',
          description: '用户原始需求 + 上传文件说明'
        },
        aiTodoList: {
          type: 'array',
          items: { type: 'string' },
          description: 'AI的执行计划清单(即使只是查看也要列出)'
        },
        aiSummary: {
          type: 'string',
          description: 'AI的操作总结(3-5句话,包括解释、分析等)'
        },
        fileOperations: {
          type: 'array',
          items: { type: 'string' },
          description: '文件操作总结,格式:"动作 文件路径 - 说明"(可为空)'
        },
        title: {
          type: 'string',
          description: '对话标题(可选)'
        },
        tags: {
          type: 'array',
          items: { type: 'string' },
          description: '标签数组(可选)'
        },
        project: {
          type: 'string',
          description: '项目名(可选,自动检测)'
        }
      },
      required: ['userRequest', 'aiTodoList', 'aiSummary']
    }
  },

• src/index.ts:161-162 (registration)

  Dispatch handler in CallToolRequest that routes 'log_conversation' calls to the ConversationLogger instance.

  case 'log_conversation':
    return await this.conversationLogger.logConversation(args || {});

• src/validation/schemas.ts:119-126 (schema)

  Validator function for log_conversation parameters using the Zod schema, throws error on invalid input.

  export const validateLogConversation = (data: unknown): LogConversationParams => {
    const result = LogConversationSchema.safeParse(data);
    if (!result.success) {
      const errorMessages = result.error.errors.map(err => `${err.path.join('.')}: ${err.message}`).join('; ');
      throw new Error(`${CONSTANTS.ERROR_MESSAGES.INVALID_PARAMETERS}: ${errorMessages}`);
    }
    return result.data as LogConversationParams;
  };