feat: Add GEMINI.md and update README files

2025-12-18 13:21:08 +08:00 · 2025-12-18 13:21:08 +08:00 · fe06ba9b1b
parent 275bc9cc3b
commit fe06ba9b1b
32 changed files with 2117 additions and 0 deletions
--- a/GEMINI.md
+++ b/GEMINI.md
@ -0,0 +1,53 @@
+<!--
+-------------------------------------------------------------------------------
+  AI Assistant Context (GEMINI.md)
+-------------------------------------------------------------------------------
+
+This document provides the necessary context for the AI assistant (Gemini) to effectively collaborate on the "Vibe Coding" project.
+
+-->
+
+<div align="center">
+
+# Vibe Coding Guide - AI Assistant Context
+
+</div>
+
+## 🚀 Project Overview
+
+**Vibe Coding** is a comprehensive guide and workflow for AI-assisted pair programming. The goal is to provide a structured and efficient way to turn ideas into reality by leveraging the power of AI. This project is not about a specific programming language or framework, but rather a methodology that can be applied to any software development project.
+
+The core of the project is a collection of documents, prompts, and tools that guide the developer and the AI assistant through the entire development process, from initial conception to final implementation.
+
+## 的核心理念 (Core Philosophy)
+
+The "Vibe Coding" methodology is based on the following key principles:
+
+*   **Planning is everything:** A detailed implementation plan is created before any code is written. This plan is then executed step-by-step, with each step being tested and verified before moving on to the next.
+*   **Modularization:** The project is broken down into small, manageable modules that can be developed and tested independently.
+*   **Context is king:** The AI assistant is provided with a "memory-bank" of all the relevant project documents, such as the game design document, tech stack, and implementation plan. This ensures that the AI has a deep understanding of the project's context and can provide accurate and relevant assistance.
+*   **AI as a partner:** The AI assistant is not just a code generator, but a true partner in the development process. The developer and the AI work together, with the developer providing the high-level guidance and the AI providing the low-level implementation details.
+
+## 📂 Folder Structure
+
+The most important files and directories in this project are:
+
+*   `README.md`: The main entry point for the project, providing an overview and links to all the other resources.
+*   `i18n/`: Contains the internationalization files for the project, with subdirectories for each supported language.
+*   `i18n/en/`: The English version of the project documentation.
+*   `i18n/zh/`: The Chinese version of the project documentation.
+*   `libs/`: Contains common library code that can be used across different projects.
+*   `prompts/`: A collection of prompts for different stages of the development process.
+*   `skills/`: A collection of reusable skills that can be used to extend the functionality of the AI assistant.
+
+## 🤖 AI Assistant's Role
+
+As the AI assistant for this project, you are expected to:
+
+*   **Be a true partner:** Work collaboratively with the developer to achieve the project's goals.
+*   **Be proactive:** Ask clarifying questions and provide suggestions to improve the project.
+*   **Be a good communicator:** Clearly explain your reasoning and provide detailed explanations of your work.
+*   **Be a good learner:** Continuously learn from your interactions with the developer and the project's context.
+*   **Follow the "Vibe Coding" methodology:** Adhere to the principles of plan-driven development, modularization, and context-awareness.
+
+By following these guidelines, you will be able to provide the best possible assistance to the developer and help them turn their ideas into reality.
--- a/README.md
+++ b/README.md
@ -303,6 +303,7 @@
 ### 项目内部文档

 *   [**胶水编程 (Glue Coding)**](./i18n/zh/documents/胶水编程/): 软件工程的圣杯与银弹，Vibe Coding 的终极进化形态。
+*   [**Chat Vault**](./libs/external/chat-vault/): AI 聊天记录保存工具，支持 Codex/Kiro/Gemini/Claude CLI。
 *   [**prompts-library 工具说明**](./libs/external/prompts-library/): 支持 Excel 与 Markdown 格式互转，包含数百个精选提示词。
 *   [**coding_prompts 集合**](./i18n/zh/prompts/coding_prompts/): 适用于 Vibe Coding 流程的专用提示词。
 *   [**系统提示词构建原则**](./i18n/zh/documents/方法论与原则/系统提示词构建原则.md): 构建高效 AI 系统提示词的综合指南。
--- a/i18n/zh/README.md
+++ b/i18n/zh/README.md
@ -221,6 +221,7 @@
    *   [**在线提示词数据库**](https://docs.google.com/spreadsheets/d/1ngoQOhJqdguwNAilCl1joNwTje7FWWN9WiI2bo5VhpU/edit?gid=2093180351#gid=2093180351&range=A1): 包含数百个适用于各场景的用户及系统提示词的在线表格。
    *   [**第三方系统提示词仓库**](https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools): 汇集了多种 AI 工具的系统提示词。
 *   **项目内部文档**:
+    *   [**Chat Vault**](../../libs/external/chat-vault/): AI 聊天记录保存工具，支持 Codex/Kiro/Gemini/Claude CLI。
    *   [**prompts-library 工具说明**](./libs/external/prompts-library/): 该工具支持在 Excel 和 Markdown 格式之间转换提示词，并包含数百个精选提示词。
    *   [**coding_prompts 集合**](./i18n/zh/prompts/coding_prompts/): 适用于 Vibe Coding 流程的专用提示词。
    *   [**系统提示词构建原则**](./documents/方法论与原则/系统提示词构建原则.md): 关于如何构建高效、可靠的 AI 系统提示词的综合指南。
--- a/libs/README.md
+++ b/libs/README.md
@ -27,8 +27,11 @@ libs/
 │   └── .gitkeep
 └── external/
    ├── README.md
+    ├── chat-vault/
    ├── prompts-library/
+    ├── l10n-tool/
    ├── my-nvim/
+    ├── MCPlayerTransfer/
    ├── XHS-image-to-PDF-conversion/
    └── .gitkeep
 ```
@ -57,6 +60,7 @@ libs/

 ## 常用入口

+- AI 聊天记录保存：[`external/chat-vault/`](./external/chat-vault/)（支持 Codex/Kiro/Gemini/Claude CLI）
 - 提示词批量管理：[`external/prompts-library/`](./external/prompts-library/)（配合 `../prompts/` 使用）
 - 备份工具：优先使用仓库根目录的 `backups/`（当前与 `libs/common/utils/backups/` 内容一致）

--- a/libs/external/README.md
+++ b/libs/external/README.md
@ -11,16 +11,22 @@
 ```
 libs/external/
 ├── README.md
+├── chat-vault/                      # AI 聊天记录保存工具
 ├── prompts-library/                 # 提示词库管理工具（Excel ↔ Markdown）
+├── l10n-tool/                       # 多语言翻译脚本
 ├── my-nvim/                         # Neovim 配置（含 nvim-config/）
+├── MCPlayerTransfer/                # MC 玩家迁移工具
 ├── XHS-image-to-PDF-conversion/     # 图片合并 PDF 工具
 └── .gitkeep
 ```

 ## 工具清单（入口与文档）

+- `chat-vault/`：AI 聊天记录保存工具，支持 Codex/Kiro/Gemini/Claude CLI（详见 [`chat-vault/README_CN.md`](./chat-vault/README_CN.md)）
 - `prompts-library/`：提示词 Excel ↔ Markdown 批量互转与索引生成（详见 [`prompts-library/README.md`](./prompts-library/README.md)）
+- `l10n-tool/`：多语言批量翻译脚本
 - `my-nvim/`：个人 Neovim 配置（详见 [`my-nvim/README.md`](./my-nvim/README.md)）
+- `MCPlayerTransfer/`：MC 玩家迁移工具
 - `XHS-image-to-PDF-conversion/`：图片合并 PDF（详见 [`XHS-image-to-PDF-conversion/README.md`](./XHS-image-to-PDF-conversion/README.md)）

 ## 新增外部工具（最小清单）
--- a/libs/external/chat-vault/.env.example
+++ b/libs/external/chat-vault/.env.example
@ -0,0 +1,11 @@
+# AI Chat Converter Configuration (Optional)
+# Default: Auto-detect paths, no configuration needed
+
+# Custom paths (comma-separated for multiple)
+# CODEX_PATHS=~/.codex/sessions
+# KIRO_PATHS=~/.local/share/kiro-cli
+# GEMINI_PATHS=~/.gemini/tmp
+# CLAUDE_PATHS=~/.claude
+
+# WSL paths also supported
+# CODEX_PATHS=\\wsl.localhost\Ubuntu\home\user\.codex\sessions
--- a/libs/external/chat-vault/.gitignore
+++ b/libs/external/chat-vault/.gitignore
@ -0,0 +1,28 @@
+# Python
+__pycache__/
+*.py[cod]
+*.so
+*.egg-info/
+dist/
+build/
+*.spec
+
+# Output
+output/
+*.db
+*.sqlite3
+*.log
+
+# Environment
+.env
+.venv/
+venv/
+
+# IDE
+.vscode/
+.idea/
+*.swp
+
+# OS
+.DS_Store
+Thumbs.db
--- a/libs/external/chat-vault/LICENSE
+++ b/libs/external/chat-vault/LICENSE
@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2024
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
--- a/libs/external/chat-vault/README.md
+++ b/libs/external/chat-vault/README.md
@ -0,0 +1,318 @@
+<div align="center">
+
+# 🔐 Chat Vault
+
+**One tool to save ALL your AI chat history**
+
+[![Python](https://img.shields.io/badge/Python-3.8+-blue.svg)](https://python.org)
+[![License](https://img.shields.io/badge/License-MIT-green.svg)](LICENSE)
+[![Platform](https://img.shields.io/badge/Platform-Linux%20%7C%20macOS%20%7C%20Windows-lightgrey.svg)]()
+[![PRs Welcome](https://img.shields.io/badge/PRs-welcome-brightgreen.svg)]()
+
+[English](README.md) | [中文](README_CN.md)
+
+[✨ Features](#-features) •
+[🚀 Quick Start](#-quick-start) •
+[📋 Commands](#-commands) •
+[📁 Project Structure](#-project-structure) •
+[❓ FAQ](#-faq)
+
+[📞 Contact](#-contact) •
+[✨ Support](#-support) •
+[🤝 Contributing](#-contributing)
+
+AI-powered docs: [zread.ai/tukuaiai/chat-vault](https://zread.ai/tukuaiai/chat-vault)
+
+> 📦 This tool is part of [vibe-coding-cn](https://github.com/tukuaiai/vibe-coding-cn) - A comprehensive Vibe Coding guide
+
+</div>
+
+---
+
+## ✨ Features
+
+<table>
+<tr>
+<td>🔄 <b>Multi-CLI</b></td>
+<td>Codex, Kiro, Gemini, Claude - all supported</td>
+</tr>
+<tr>
+<td>⚡ <b>Real-time</b></td>
+<td>Watch mode with system-level file monitoring</td>
+</tr>
+<tr>
+<td>🔢 <b>Token Stats</b></td>
+<td>Accurate counting using tiktoken (cl100k_base)</td>
+</tr>
+<tr>
+<td>🔍 <b>Search</b></td>
+<td>Find any conversation instantly</td>
+</tr>
+<tr>
+<td>📤 <b>Export</b></td>
+<td>JSON or CSV, your choice</td>
+</tr>
+<tr>
+<td>🚀 <b>Zero Config</b></td>
+<td>Auto-detects paths, just run it</td>
+</tr>
+</table>
+
+---
+
+## 🏗️ Architecture
+
+```mermaid
+graph LR
+    subgraph Sources
+        A[~/.codex] 
+        B[~/.kiro]
+        C[~/.gemini]
+        D[~/.claude]
+    end
+    
+    subgraph Chat Vault
+        E[Watcher]
+        F[Parsers]
+        G[Storage]
+    end
+    
+    subgraph Output
+        H[(SQLite DB)]
+    end
+    
+    A --> E
+    B --> E
+    C --> E
+    D --> E
+    E --> F
+    F --> G
+    G --> H
+```
+
+---
+
+## 🔄 How It Works
+
+```mermaid
+sequenceDiagram
+    participant User
+    participant CLI as AI CLI (Codex/Kiro/...)
+    participant Watcher
+    participant Parser
+    participant DB as SQLite
+
+    User->>CLI: Chat with AI
+    CLI->>CLI: Save to local file
+    Watcher->>Watcher: Detect file change
+    Watcher->>Parser: Parse new content
+    Parser->>DB: Upsert session
+    DB-->>User: Query anytime
+```
+
+---
+
+## 🚀 Quick Start
+
+### 30 Seconds Setup
+
+```bash
+# Clone
+git clone https://github.com/tukuaiai/vibe-coding-cn.git
+cd vibe-coding-cn/libs/external/chat-vault
+
+# Run (auto-installs dependencies)
+./start.sh        # Linux/macOS
+start.bat         # Windows
+```
+
+**That's it!** 🎉
+
+---
+
+## 📊 Example Output
+
+```
+==================================================
+AI 聊天记录 → 集中存储
+==================================================
+数据库: ./output/chat_history.db
+
+[Codex] 新增:1241 更新:0 跳过:0 错误:0
+[Kiro] 新增:21 更新:0 跳过:0 错误:0
+[Gemini] 新增:332 更新:0 跳过:0 错误:0
+[Claude] 新增:168 更新:0 跳过:0 错误:0
+
+==================================================
+总计: 1762 会话, 40000+ 消息
+✓ 同步完成!
+
+=== Token 统计 (tiktoken) ===
+  codex: 11,659,952 tokens
+  kiro: 26,337 tokens
+  gemini: 3,195,821 tokens
+  claude: 29,725 tokens
+  总计: 14,911,835 tokens
+```
+
+---
+
+## 📋 Commands
+
+| Command | Description |
+|---------|-------------|
+| `python src/main.py` | Sync once |
+| `python src/main.py -w` | Watch mode (real-time) |
+| `python src/main.py --stats` | Show statistics |
+| `python src/main.py --search "keyword"` | Search messages |
+| `python src/main.py --export json` | Export to JSON |
+| `python src/main.py --export csv --source codex` | Export specific source |
+| `python src/main.py --prune` | Clean orphaned records |
+
+---
+
+## 📁 Project Structure
+
+```
+chat-vault/
+├── 🚀 start.sh / start.bat    # One-click start
+├── 📦 build.py                # Build standalone exe
+├── 📂 src/
+│   ├── main.py                # CLI entry
+│   ├── config.py              # Auto-detection
+│   ├── storage.py             # SQLite + tiktoken
+│   ├── watcher.py             # File monitoring
+│   └── parsers/               # CLI parsers
+├── 📂 docs/
+│   ├── AI_PROMPT.md           # AI assistant guide
+│   └── schema.md              # Database schema
+└── 📂 output/
+    ├── chat_history.db        # Your database
+    └── logs/                   # Sync logs
+```
+
+---
+
+## 🗄️ Database Schema
+
+```mermaid
+erDiagram
+    sessions {
+        TEXT file_path PK
+        TEXT session_id
+        TEXT source
+        TEXT cwd
+        TEXT messages
+        INTEGER file_mtime
+        TEXT start_time
+        INTEGER token_count
+    }
+    
+    meta {
+        TEXT key PK
+        TEXT value
+    }
+    
+    meta_codex {
+        TEXT key PK
+        TEXT value
+    }
+```
+
+---
+
+## 🤖 For AI Assistants
+
+Send [docs/AI_PROMPT.md](docs/AI_PROMPT.md) to your AI assistant for:
+- SQL query examples
+- Python code snippets
+- Task guidance
+
+---
+
+## ❓ FAQ
+
+<details>
+<summary><b>Do I need to configure anything?</b></summary>
+
+No. Auto-detects `~/.codex`, `~/.kiro`, `~/.gemini`, `~/.claude`
+</details>
+
+<details>
+<summary><b>Does it work with WSL?</b></summary>
+
+Yes! Paths like `\\wsl.localhost\Ubuntu\...` are supported
+</details>
+
+<details>
+<summary><b>How do I view the database?</b></summary>
+
+Use [DB Browser for SQLite](https://sqlitebrowser.org/) or any SQLite tool
+</details>
+
+<details>
+<summary><b>Is my data safe?</b></summary>
+
+Yes. We only READ from AI tools, never modify original files
+</details>
+
+---
+
+## 📞 Contact
+
+- **GitHub**: [tukuaiai](https://github.com/tukuaiai)
+- **Twitter / X**: [123olp](https://x.com/123olp)
+- **Telegram**: [@desci0](https://t.me/desci0)
+- **Telegram Group**: [glue_coding](https://t.me/glue_coding)
+- **Telegram Channel**: [tradecat_ai_channel](https://t.me/tradecat_ai_channel)
+- **Email**: tukuai.ai@gmail.com
+
+---
+
+## ✨ Support
+
+If this project helped you, consider supporting:
+
+- **Binance UID**: `572155580`
+- **Tron (TRC20)**: `TQtBXCSTwLFHjBqTS4rNUp7ufiGx51BRey`
+- **Solana**: `HjYhozVf9AQmfv7yv79xSNs6uaEU5oUk2USasYQfUYau`
+- **Ethereum (ERC20)**: `0xa396923a71ee7D9480b346a17dDeEb2c0C287BBC`
+- **BNB Smart Chain (BEP20)**: `0xa396923a71ee7D9480b346a17dDeEb2c0C287BBC`
+- **Bitcoin**: `bc1plslluj3zq3snpnnczplu7ywf37h89dyudqua04pz4txwh8z5z5vsre7nlm`
+- **Sui**: `0xb720c98a48c77f2d49d375932b2867e793029e6337f1562522640e4f84203d2e`
+
+---
+
+## 🤝 Contributing
+
+We welcome all contributions! Feel free to open an [Issue](https://github.com/tukuaiai/vibe-coding-cn/issues) or submit a [Pull Request](https://github.com/tukuaiai/vibe-coding-cn/pulls).
+
+---
+
+## 📄 License
+
+[MIT](LICENSE) - Do whatever you want with it.
+
+---
+
+<div align="center">
+
+**If this helped you, give it a ⭐!**
+
+## Star History
+
+<a href="https://www.star-history.com/#tukuaiai/vibe-coding-cn&type=Date">
+ <picture>
+   <source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/svg?repos=tukuaiai/vibe-coding-cn&type=Date&theme=dark" />
+   <source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/svg?repos=tukuaiai/vibe-coding-cn&type=Date" />
+   <img alt="Star History Chart" src="https://api.star-history.com/svg?repos=tukuaiai/vibe-coding-cn&type=Date" />
+ </picture>
+</a>
+
+---
+
+**Made with ❤️ by [tukuaiai](https://github.com/tukuaiai)**
+
+[⬆ Back to Top](#-chat-vault)
+
+</div>
--- a/libs/external/chat-vault/README_CN.md
+++ b/libs/external/chat-vault/README_CN.md
@ -0,0 +1,311 @@
+<div align="center">
+
+# 🔐 Chat Vault
+
+**一个工具保存你所有的 AI 聊天记录**
+
+[![Python](https://img.shields.io/badge/Python-3.8+-blue.svg)](https://python.org)
+[![License](https://img.shields.io/badge/License-MIT-green.svg)](LICENSE)
+[![Platform](https://img.shields.io/badge/Platform-Linux%20%7C%20macOS%20%7C%20Windows-lightgrey.svg)]()
+[![PRs Welcome](https://img.shields.io/badge/PRs-welcome-brightgreen.svg)]()
+
+[English](README.md) | [中文](README_CN.md)
+
+[✨ 功能特性](#-功能特性) •
+[🚀 快速开始](#-30-秒快速开始) •
+[📋 命令一览](#-命令一览) •
+[📁 项目结构](#-项目结构) •
+[❓ 常见问题](#-常见问题)
+
+[📞 联系方式](#-联系方式) •
+[✨ 支持项目](#-支持项目) •
+[🤝 参与贡献](#-参与贡献)
+
+AI 解读文档: [zread.ai/tukuaiai/chat-vault](https://zread.ai/tukuaiai/chat-vault)
+
+> 📦 本工具是 [vibe-coding-cn](https://github.com/tukuaiai/vibe-coding-cn) 的一部分 - 一份全面的 Vibe Coding 指南
+
+</div>
+
+---
+
+## ✨ 功能特性
+
+<table>
+<tr>
+<td>🔄 <b>多 CLI 支持</b></td>
+<td>Codex、Kiro、Gemini、Claude 全都行</td>
+</tr>
+<tr>
+<td>⚡ <b>实时同步</b></td>
+<td>系统级文件监控，聊完自动保存</td>
+</tr>
+<tr>
+<td>🔢 <b>Token 统计</b></td>
+<td>tiktoken 精确计算，知道你用了多少</td>
+</tr>
+<tr>
+<td>🔍 <b>搜索</b></td>
+<td>秒找任何对话</td>
+</tr>
+<tr>
+<td>📤 <b>导出</b></td>
+<td>JSON 或 CSV，随你选</td>
+</tr>
+<tr>
+<td>🚀 <b>零配置</b></td>
+<td>自动检测路径，开箱即用</td>
+</tr>
+</table>
+
+---
+
+## 🏗️ 架构图
+
+```mermaid
+graph LR
+    subgraph 数据来源
+        A[~/.codex] 
+        B[~/.kiro]
+        C[~/.gemini]
+        D[~/.claude]
+    end
+    
+    subgraph Chat Vault
+        E[监控器]
+        F[解析器]
+        G[存储层]
+    end
+    
+    subgraph 输出
+        H[(SQLite 数据库)]
+    end
+    
+    A --> E
+    B --> E
+    C --> E
+    D --> E
+    E --> F
+    F --> G
+    G --> H
+```
+
+---
+
+## 🔄 工作流程
+
+```mermaid
+sequenceDiagram
+    participant 用户
+    participant CLI as AI CLI (Codex/Kiro/...)
+    participant 监控器
+    participant 解析器
+    participant DB as SQLite
+
+    用户->>CLI: 和 AI 聊天
+    CLI->>CLI: 保存到本地文件
+    监控器->>监控器: 检测文件变化
+    监控器->>解析器: 解析新内容
+    解析器->>DB: 写入数据库
+    DB-->>用户: 随时查询
+```
+
+---
+
+## 🚀 30 秒快速开始
+
+```bash
+# 下载
+git clone https://github.com/tukuaiai/vibe-coding-cn.git
+cd vibe-coding-cn/libs/external/chat-vault
+
+# 运行（自动安装依赖）
+./start.sh        # Linux/macOS
+start.bat         # Windows（双击）
+```
+
+**搞定！** 🎉
+
+---
+
+## 📊 运行效果
+
+```
+==================================================
+AI 聊天记录 → 集中存储
+==================================================
+数据库: ./output/chat_history.db
+
+[Codex] 新增:1241 更新:0 跳过:0 错误:0
+[Kiro] 新增:21 更新:0 跳过:0 错误:0
+[Gemini] 新增:332 更新:0 跳过:0 错误:0
+[Claude] 新增:168 更新:0 跳过:0 错误:0
+
+==================================================
+总计: 1762 会话, 40000+ 消息
+✓ 同步完成!
+
+=== Token 统计 (tiktoken) ===
+  codex: 11,659,952 tokens
+  kiro: 26,337 tokens
+  gemini: 3,195,821 tokens
+  claude: 29,725 tokens
+  总计: 14,911,835 tokens
+```
+
+---
+
+## 📋 命令一览
+
+| 命令 | 说明 |
+|------|------|
+| `python src/main.py` | 同步一次 |
+| `python src/main.py -w` | 实时监控（推荐） |
+| `python src/main.py --stats` | 查看统计 |
+| `python src/main.py --search "关键词"` | 搜索消息 |
+| `python src/main.py --export json` | 导出 JSON |
+| `python src/main.py --export csv --source codex` | 导出指定来源 |
+| `python src/main.py --prune` | 清理孤立记录 |
+
+---
+
+## 📁 项目结构
+
+```
+chat-vault/
+├── 🚀 start.sh / start.bat    # 一键启动
+├── 📦 build.py                # 打包脚本
+├── 📂 src/
+│   ├── main.py                # 主程序
+│   ├── config.py              # 配置检测
+│   ├── storage.py             # SQLite + tiktoken
+│   ├── watcher.py             # 文件监控
+│   └── parsers/               # 各 CLI 解析器
+├── 📂 docs/
+│   ├── AI_PROMPT.md           # AI 助手指南
+│   └── schema.md              # 数据库结构
+└── 📂 output/
+    ├── chat_history.db        # 你的数据库
+    └── logs/                   # 日志
+```
+
+---
+
+## 🗄️ 数据库结构
+
+```mermaid
+erDiagram
+    sessions {
+        TEXT file_path PK "文件路径"
+        TEXT session_id "会话ID"
+        TEXT source "来源"
+        TEXT cwd "工作目录"
+        TEXT messages "消息JSON"
+        INTEGER file_mtime "修改时间"
+        TEXT start_time "开始时间"
+        INTEGER token_count "Token数"
+    }
+    
+    meta {
+        TEXT key PK
+        TEXT value
+    }
+```
+
+---
+
+## 🤖 让 AI 帮你查数据库
+
+把 [docs/AI_PROMPT.md](docs/AI_PROMPT.md) 发给 AI 助手，它就知道：
+- 怎么写 SQL 查询
+- 怎么用 Python 分析
+- 怎么帮你找对话
+
+---
+
+## ❓ 常见问题
+
+<details>
+<summary><b>需要配置什么吗？</b></summary>
+
+不用。自动检测 `~/.codex`、`~/.kiro`、`~/.gemini`、`~/.claude`
+</details>
+
+<details>
+<summary><b>WSL 能用吗？</b></summary>
+
+能！`\\wsl.localhost\Ubuntu\...` 这种路径也支持
+</details>
+
+<details>
+<summary><b>怎么看数据库？</b></summary>
+
+用 [DB Browser for SQLite](https://sqlitebrowser.org/) 或任何 SQLite 工具
+</details>
+
+<details>
+<summary><b>会不会搞坏我的数据？</b></summary>
+
+不会。只读取，从不修改原始文件
+</details>
+
+---
+
+## 📞 联系方式
+
+- **GitHub**: [tukuaiai](https://github.com/tukuaiai)
+- **Twitter / X**: [123olp](https://x.com/123olp)
+- **Telegram**: [@desci0](https://t.me/desci0)
+- **Telegram 交流群**: [glue_coding](https://t.me/glue_coding)
+- **Telegram 频道**: [tradecat_ai_channel](https://t.me/tradecat_ai_channel)
+- **邮箱**: tukuai.ai@gmail.com
+
+---
+
+## ✨ 支持项目
+
+如果这个项目帮到你了，考虑支持一下：
+
+- **币安 UID**: `572155580`
+- **Tron (TRC20)**: `TQtBXCSTwLFHjBqTS4rNUp7ufiGx51BRey`
+- **Solana**: `HjYhozVf9AQmfv7yv79xSNs6uaEU5oUk2USasYQfUYau`
+- **Ethereum (ERC20)**: `0xa396923a71ee7D9480b346a17dDeEb2c0C287BBC`
+- **BNB Smart Chain (BEP20)**: `0xa396923a71ee7D9480b346a17dDeEb2c0C287BBC`
+- **Bitcoin**: `bc1plslluj3zq3snpnnczplu7ywf37h89dyudqua04pz4txwh8z5z5vsre7nlm`
+- **Sui**: `0xb720c98a48c77f2d49d375932b2867e793029e6337f1562522640e4f84203d2e`
+
+---
+
+## 🤝 参与贡献
+
+欢迎各种形式的贡献！随时开启一个 [Issue](https://github.com/tukuaiai/vibe-coding-cn/issues) 或提交 [Pull Request](https://github.com/tukuaiai/vibe-coding-cn/pulls)。
+
+---
+
+## 📄 开源协议
+
+[MIT](LICENSE) - 随便用，不用管我
+
+---
+
+<div align="center">
+
+**如果帮到你了，点个 ⭐ 呗！**
+
+## Star History
+
+<a href="https://www.star-history.com/#tukuaiai/vibe-coding-cn&type=Date">
+ <picture>
+   <source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/svg?repos=tukuaiai/vibe-coding-cn&type=Date&theme=dark" />
+   <source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/svg?repos=tukuaiai/vibe-coding-cn&type=Date" />
+   <img alt="Star History Chart" src="https://api.star-history.com/svg?repos=tukuaiai/vibe-coding-cn&type=Date" />
+ </picture>
+</a>
+
+---
+
+**Made with ❤️ by [tukuaiai](https://github.com/tukuaiai)**
+
+[⬆ 返回顶部](#-chat-vault)
+
+</div>
--- a/libs/external/chat-vault/build.bat
+++ b/libs/external/chat-vault/build.bat
@ -0,0 +1,15 @@
+@echo off
+echo 安装打包工具...
+pip install pyinstaller -q
+
+echo 开始打包...
+pyinstaller --onefile --name ai-chat-converter ^
+    --add-data "src;src" ^
+    --hidden-import tiktoken_ext.openai_public ^
+    --hidden-import tiktoken_ext ^
+    --collect-data tiktoken ^
+    src/main.py
+
+echo.
+echo 完成! 输出: dist\ai-chat-converter.exe
+pause
--- a/libs/external/chat-vault/build.py
+++ b/libs/external/chat-vault/build.py
@ -0,0 +1,40 @@
+#!/usr/bin/env python3
+"""打包脚本 - 生成独立可执行文件"""
+import subprocess
+import sys
+import os
+import shutil
+
+def main():
+    os.chdir(os.path.dirname(os.path.abspath(__file__)))
+    
+    for d in ['build', 'dist']:
+        if os.path.exists(d):
+            shutil.rmtree(d)
+    
+    print("开始打包...")
+    
+    sep = ";" if sys.platform == "win32" else ":"
+    
+    cmd = [
+        sys.executable, "-m", "PyInstaller",
+        "--onefile",
+        "--name", "ai-chat-converter",
+        f"--add-data=src{sep}src",
+        "--hidden-import", "tiktoken_ext.openai_public",
+        "--hidden-import", "tiktoken_ext",
+        "--hidden-import", "dotenv",
+        "--collect-data", "tiktoken",
+        "--collect-all", "watchdog",
+        "--collect-all", "dotenv",
+        "src/main.py"
+    ]
+    
+    subprocess.run(cmd, check=True)
+    
+    exe = "dist/ai-chat-converter.exe" if sys.platform == "win32" else "dist/ai-chat-converter"
+    size = os.path.getsize(exe) / 1024 / 1024
+    print(f"\n✓ 打包完成: {exe} ({size:.1f} MB)")
+
+if __name__ == "__main__":
+    main()
--- a/libs/external/chat-vault/docs/AI_PROMPT.md
+++ b/libs/external/chat-vault/docs/AI_PROMPT.md
@ -0,0 +1,231 @@
+# AI Chat Converter - AI 助手完全指南
+
+> **把这个文档发给 AI 助手，它就知道怎么帮你用这个工具了**
+
+---
+
+## 🎯 这是什么？
+
+一个把 Codex、Kiro、Gemini、Claude 的聊天记录全部存到一个 SQLite 数据库的工具。
+
+**数据库位置**: `项目目录/output/chat_history.db`
+
+---
+
+## 🚀 怎么启动？
+
+### 方式一：双击启动（推荐）
+```bash
+./start.sh          # Linux/macOS
+start.bat           # Windows（双击）
+```
+
+### 方式二：命令行
+```bash
+cd ai-chat-converter
+python src/main.py --watch   # 持续监控（推荐）
+python src/main.py           # 同步一次就退出
+```
+
+### 方式三：后台运行
+```bash
+nohup ./start.sh > /dev/null 2>&1 &
+```
+
+---
+
+## 📊 数据库长啥样？
+
+### 主表：sessions
+
+| 字段 | 说明 | 例子 |
+|------|------|------|
+| file_path | 主键，文件路径 | `/home/user/.codex/sessions/xxx.jsonl` |
+| session_id | 会话ID | `019b2164-168c-7133-9b1f-5d24fea1d3e1` |
+| source | 来源 | `codex` / `kiro` / `gemini` / `claude` |
+| cwd | 工作目录 | `/home/user/projects/myapp` |
+| messages | 消息内容（JSON） | `[{"time":"...", "role":"user", "content":"..."}]` |
+| start_time | 开始时间 | `2025-12-18T10:30:00` |
+| token_count | Token 数量 | `1234` |
+
+---
+
+## 🔍 常用查询（直接复制用）
+
+### 1. 看看有多少数据
+
+```sql
+SELECT source, COUNT(*) as 会话数, SUM(token_count) as Token总数
+FROM sessions 
+GROUP BY source;
+```
+
+### 2. 最近的 10 个会话
+
+```sql
+SELECT session_id, source, cwd, start_time, token_count
+FROM sessions
+ORDER BY start_time DESC
+LIMIT 10;
+```
+
+### 3. 搜索包含某个词的对话
+
+```sql
+SELECT session_id, source, cwd, start_time
+FROM sessions
+WHERE messages LIKE '%要搜索的词%'
+ORDER BY start_time DESC
+LIMIT 20;
+```
+
+### 4. 查某个项目的所有对话
+
+```sql
+SELECT session_id, source, start_time, token_count
+FROM sessions
+WHERE cwd LIKE '%项目名%'
+ORDER BY start_time;
+```
+
+### 5. 看某个会话的完整内容
+
+```sql
+SELECT messages FROM sessions WHERE session_id = '会话ID';
+```
+
+### 6. 统计每天用了多少 Token
+
+```sql
+SELECT 
+    date(start_time) as 日期,
+    SUM(token_count) as Token数
+FROM sessions
+GROUP BY 日期
+ORDER BY 日期 DESC
+LIMIT 7;
+```
+
+### 7. 统计每个来源的 Token
+
+```sql
+SELECT source, SUM(token_count) as tokens
+FROM sessions
+GROUP BY source
+ORDER BY tokens DESC;
+```
+
+---
+
+## 💻 命令行用法
+
+| 命令 | 干啥的 |
+|------|--------|
+| `python src/main.py` | 同步一次 |
+| `python src/main.py -w` | 持续监控（推荐） |
+| `python src/main.py --stats` | 看统计信息 |
+| `python src/main.py --search "关键词"` | 搜索 |
+| `python src/main.py --export json` | 导出 JSON |
+| `python src/main.py --export csv` | 导出 CSV |
+| `python src/main.py --prune` | 清理已删除文件的记录 |
+
+---
+
+## 🐍 用 Python 查询
+
+```python
+import sqlite3
+import json
+
+# 连接数据库
+db = sqlite3.connect('output/chat_history.db')
+
+# 查所有 Codex 会话
+for row in db.execute("SELECT session_id, cwd, token_count FROM sessions WHERE source='codex'"):
+    print(f"{row[0]}: {row[2]} tokens - {row[1]}")
+
+# 搜索包含 "python" 的对话
+for row in db.execute("SELECT session_id, source FROM sessions WHERE messages LIKE '%python%'"):
+    print(f"[{row[1]}] {row[0]}")
+
+# 获取某个会话的消息
+row = db.execute("SELECT messages FROM sessions WHERE session_id=?", ('会话ID',)).fetchone()
+if row:
+    messages = json.loads(row[0])
+    for msg in messages:
+        print(f"{msg['role']}: {msg['content'][:100]}...")
+```
+
+---
+
+## 📁 文件在哪？
+
+```
+ai-chat-converter/
+├── start.sh              ← 双击这个启动
+├── output/
+│   ├── chat_history.db   ← 数据库在这
+│   └── logs/             ← 日志在这
+└── src/
+    └── main.py           ← 主程序
+```
+
+---
+
+## ❓ AI 助手任务示例
+
+当用户说这些话时，你应该这样做：
+
+| 用户说 | 你做 |
+|--------|------|
+| "帮我查最近的对话" | 执行最近会话 SQL |
+| "搜索关于 Python 的讨论" | 用 `--search` 或 SQL 搜索 |
+| "这个月用了多少 Token" | 执行 Token 统计 SQL |
+| "导出所有 Codex 记录" | `python src/main.py --export json --source codex` |
+| "启动监控" | `./start.sh` 或 `python src/main.py -w` |
+| "数据库在哪" | `output/chat_history.db` |
+
+---
+
+## 🔧 出问题了？
+
+### 问题：找不到数据库
+```bash
+# 先运行一次同步
+python src/main.py
+```
+
+### 问题：依赖没装
+```bash
+pip install -r requirements.txt
+```
+
+### 问题：权限不够
+```bash
+chmod +x start.sh
+```
+
+---
+
+## 📊 消息格式
+
+数据库里的 `messages` 字段是 JSON 数组：
+
+```json
+[
+  {
+    "time": "2025-12-18T10:30:00",
+    "role": "user",
+    "content": "帮我写个 Python 脚本"
+  },
+  {
+    "time": "2025-12-18T10:30:05",
+    "role": "ai",
+    "content": "好的，这是一个简单的脚本..."
+  }
+]
+```
+
+- `role`: `user`（用户）或 `ai`（AI 回复）
+- `time`: ISO 格式时间
+- `content`: 消息内容
--- a/libs/external/chat-vault/docs/roadmap.md
+++ b/libs/external/chat-vault/docs/roadmap.md
@ -0,0 +1,15 @@
+# Roadmap
+
+## v1.0 ✅ Core
+- [x] Multi-CLI support (Codex/Kiro/Gemini/Claude)
+- [x] Auto path detection
+- [x] SQLite storage
+- [x] Incremental sync
+- [x] Cross-platform watch mode (watchdog)
+- [x] Token counting (tiktoken)
+
+## Future
+- [ ] Web UI
+- [ ] API server mode
+- [ ] Vector storage (RAG)
+- [ ] Cross-AI context sharing
--- a/libs/external/chat-vault/docs/schema.md
+++ b/libs/external/chat-vault/docs/schema.md
@ -0,0 +1,49 @@
+# 数据库结构 (v5)
+
+**位置**: `项目目录/output/chat_history.db`
+
+## sessions 表（主表）
+
+| 字段 | 类型 | 说明 |
+|------|------|------|
+| file_path | TEXT | 主键，源文件路径 |
+| session_id | TEXT | 会话 ID |
+| source | TEXT | 来源: codex/kiro/gemini/claude |
+| cwd | TEXT | 工作目录 |
+| messages | TEXT | JSON 数组 |
+| file_mtime | INTEGER | 文件修改时间戳 |
+| start_time | TEXT | 会话开始时间 |
+| token_count | INTEGER | Token 数量 |
+
+**索引**: `idx_source`, `idx_session_id`, `idx_start_time`
+
+## meta 表（全局统计）
+
+| key | 说明 |
+|-----|------|
+| schema_version | 数据库版本 (5) |
+| total_sessions | 总会话数 |
+| total_messages | 总消息数 |
+| total_tokens | 总 Token 数 |
+| last_sync | 最后同步时间 |
+
+## meta_{cli} 表（各 CLI 统计）
+
+每个 CLI 独立的元信息表：`meta_codex`, `meta_kiro`, `meta_gemini`, `meta_claude`
+
+| key | 说明 |
+|-----|------|
+| path | 监控路径 |
+| sessions | 会话数 |
+| messages | 消息数 |
+| total_tokens | Token 总数 |
+| last_sync | 最后同步时间 |
+
+## 消息格式
+
+```json
+[
+  {"time": "2025-12-18T10:30:00", "role": "user", "content": "..."},
+  {"time": "2025-12-18T10:30:05", "role": "ai", "content": "..."}
+]
+```
--- a/libs/external/chat-vault/requirements.txt
+++ b/libs/external/chat-vault/requirements.txt
@ -0,0 +1,3 @@
+python-dotenv>=1.0.0
+watchdog>=3.0.0
+tiktoken>=0.5.0
--- a/libs/external/chat-vault/scripts/sync.sh
+++ b/libs/external/chat-vault/scripts/sync.sh
@ -0,0 +1,3 @@
+#!/bin/bash
+cd "$(dirname "$0")/../src"
+python3 main.py
--- a/libs/external/chat-vault/scripts/watch.sh
+++ b/libs/external/chat-vault/scripts/watch.sh
@ -0,0 +1,3 @@
+#!/bin/bash
+cd "$(dirname "$0")/../src"
+python3 main.py --watch
--- a/libs/external/chat-vault/src/config.py
+++ b/libs/external/chat-vault/src/config.py
@ -0,0 +1,69 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+r"""
+配置模块 - 智能路径识别
+支持: Linux 原生路径、WSL 路径 (\\wsl.localhost\Ubuntu\...)
+"""
+import os
+import re
+from dotenv import load_dotenv
+
+# 项目目录
+PROJECT_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+OUTPUT_DIR = os.path.join(PROJECT_DIR, "output")
+
+load_dotenv(os.path.join(PROJECT_DIR, ".env"))
+
+def convert_wsl_path(path: str) -> str:
+    match = re.match(r'^\\\\wsl[.\$]?[^\\]*\\[^\\]+\\(.+)$', path, re.IGNORECASE)
+    if match:
+        return '/' + match.group(1).replace('\\', '/')
+    return path
+
+def normalize_path(path: str) -> str:
+    path = path.strip()
+    path = convert_wsl_path(path)
+    return os.path.expanduser(path)
+
+def get_paths(env_key: str) -> list:
+    val = os.getenv(env_key, "")
+    if not val:
+        return []
+    return [normalize_path(p) for p in val.split(",") if p.strip()]
+
+def auto_detect_paths() -> dict:
+    home = os.path.expanduser("~")
+    kiro_db = os.path.join(home, ".local", "share", "kiro-cli")
+    candidates = {
+        "codex_paths": [os.path.join(home, ".codex", "sessions"), os.path.join(home, ".codex")],
+        "kiro_paths": [kiro_db] if os.path.exists(kiro_db) else [],
+        "gemini_paths": [os.path.join(home, ".gemini", "tmp"), os.path.join(home, ".gemini")],
+        "claude_paths": [os.path.join(home, ".claude")],
+    }
+    detected = {}
+    for key, paths in candidates.items():
+        for p in paths:
+            if os.path.exists(p):
+                detected[key] = [p]
+                break
+        if key not in detected:
+            detected[key] = []
+    return detected
+
+def load_config() -> dict:
+    auto = auto_detect_paths()
+    
+    os.makedirs(OUTPUT_DIR, exist_ok=True)
+    os.makedirs(os.path.join(OUTPUT_DIR, "logs"), exist_ok=True)
+    
+    return {
+        "codex_paths": get_paths("CODEX_PATHS") or auto.get("codex_paths", []),
+        "kiro_paths": get_paths("KIRO_PATHS") or auto.get("kiro_paths", []),
+        "gemini_paths": get_paths("GEMINI_PATHS") or auto.get("gemini_paths", []),
+        "claude_paths": get_paths("CLAUDE_PATHS") or auto.get("claude_paths", []),
+        "output_dir": OUTPUT_DIR,
+        "log_dir": os.path.join(OUTPUT_DIR, "logs"),
+        "db_path": os.path.join(OUTPUT_DIR, "chat_history.db"),
+    }
+
+CONFIG = load_config()
--- a/libs/external/chat-vault/src/logger.py
+++ b/libs/external/chat-vault/src/logger.py
@ -0,0 +1,42 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""日志模块 - 同时输出到控制台和文件"""
+import logging
+import os
+from datetime import datetime
+
+_logger = None
+
+def setup_logger(log_dir: str = None) -> logging.Logger:
+    global _logger
+    if _logger:
+        return _logger
+    
+    _logger = logging.getLogger('ai_chat_converter')
+    _logger.setLevel(logging.DEBUG)
+    _logger.handlers.clear()
+    
+    fmt = logging.Formatter('%(asctime)s [%(levelname)s] %(message)s', datefmt='%Y-%m-%d %H:%M:%S')
+    
+    # 控制台
+    ch = logging.StreamHandler()
+    ch.setLevel(logging.INFO)
+    ch.setFormatter(fmt)
+    _logger.addHandler(ch)
+    
+    # 文件
+    if log_dir:
+        os.makedirs(log_dir, exist_ok=True)
+        log_file = os.path.join(log_dir, f"sync_{datetime.now().strftime('%Y%m%d')}.log")
+        fh = logging.FileHandler(log_file, encoding='utf-8')
+        fh.setLevel(logging.DEBUG)
+        fh.setFormatter(fmt)
+        _logger.addHandler(fh)
+    
+    return _logger
+
+def get_logger() -> logging.Logger:
+    global _logger
+    if not _logger:
+        _logger = setup_logger()
+    return _logger
--- a/libs/external/chat-vault/src/main.py
+++ b/libs/external/chat-vault/src/main.py
@ -0,0 +1,319 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+AI 聊天记录集中存储工具
+
+命令:
+  python main.py              # 同步一次
+  python main.py --watch      # 持续监控
+  python main.py --prune      # 清理孤立记录
+  python main.py --stats      # 显示统计
+  python main.py --search <keyword>  # 搜索
+  python main.py --export json|csv [--source codex|kiro|gemini|claude]
+"""
+import os
+import sys
+import subprocess
+
+# 项目根目录
+PROJECT_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+VENV_DIR = os.path.join(PROJECT_DIR, '.venv')
+REQUIREMENTS = os.path.join(PROJECT_DIR, 'requirements.txt')
+
+def ensure_venv():
+    """检测并创建虚拟环境，安装依赖"""
+    # 打包版本跳过
+    if getattr(sys, 'frozen', False):
+        return
+    
+    # 已在虚拟环境中运行则跳过
+    if sys.prefix != sys.base_prefix:
+        return
+    
+    # 检查 .venv 是否存在
+    venv_python = os.path.join(VENV_DIR, 'bin', 'python') if os.name != 'nt' else os.path.join(VENV_DIR, 'Scripts', 'python.exe')
+    
+    if not os.path.exists(venv_python):
+        print("首次运行，创建虚拟环境...")
+        subprocess.run([sys.executable, '-m', 'venv', VENV_DIR], check=True)
+        print("安装依赖...")
+        pip = os.path.join(VENV_DIR, 'bin', 'pip') if os.name != 'nt' else os.path.join(VENV_DIR, 'Scripts', 'pip.exe')
+        subprocess.run([pip, 'install', '-r', REQUIREMENTS, '-q'], check=True)
+        print("环境准备完成，重新启动...\n")
+    
+    # 使用虚拟环境重新执行
+    os.execv(venv_python, [venv_python] + sys.argv)
+
+# 启动前检测虚拟环境
+ensure_venv()
+
+# 支持 PyInstaller 打包
+if getattr(sys, 'frozen', False):
+    BASE_DIR = sys._MEIPASS
+    sys.path.insert(0, os.path.join(BASE_DIR, 'src'))
+else:
+    BASE_DIR = os.path.dirname(os.path.abspath(__file__))
+
+import argparse
+from config import CONFIG
+from parsers import CodexParser, GeminiParser, ClaudeParser, KiroParser
+from storage import ChatStorage
+from logger import setup_logger, get_logger
+
+storage: ChatStorage = None
+
+def main():
+    global storage
+    
+    parser = argparse.ArgumentParser(description='AI Chat Converter')
+    parser.add_argument('-w', '--watch', action='store_true', help='持续监控模式')
+    parser.add_argument('--prune', action='store_true', help='清理孤立记录')
+    parser.add_argument('--stats', action='store_true', help='显示统计信息')
+    parser.add_argument('--search', type=str, help='搜索关键词')
+    parser.add_argument('--export', choices=['json', 'csv'], help='导出格式')
+    parser.add_argument('--source', choices=['codex', 'kiro', 'gemini', 'claude'], help='指定来源')
+    parser.add_argument('--output', type=str, help='导出文件路径')
+    args = parser.parse_args()
+    
+    # 初始化
+    setup_logger(CONFIG["log_dir"])
+    log = get_logger()
+    
+    storage = ChatStorage(CONFIG["db_path"])
+    
+    # 命令分发
+    if args.prune:
+        cmd_prune()
+    elif args.stats:
+        cmd_stats()
+    elif args.search:
+        cmd_search(args.search, args.source)
+    elif args.export:
+        cmd_export(args.export, args.source, args.output)
+    elif args.watch:
+        cmd_sync()
+        cmd_watch()
+    else:
+        cmd_sync()
+
+def cmd_sync():
+    log = get_logger()
+    log.info("=" * 50)
+    log.info("AI 聊天记录 → 集中存储")
+    log.info("=" * 50)
+    log.info(f"数据库: {CONFIG['db_path']}")
+    
+    total_added, total_updated, total_skipped, total_errors = 0, 0, 0, 0
+    
+    for cli, key, parser_cls in [
+        ('codex', 'codex_paths', lambda: CodexParser('codex')),
+        ('kiro', 'kiro_paths', KiroParser),
+        ('gemini', 'gemini_paths', GeminiParser),
+        ('claude', 'claude_paths', ClaudeParser),
+    ]:
+        paths = CONFIG.get(key, [])
+        if not paths:
+            continue
+        
+        parser = parser_cls()
+        if cli in ('claude', 'kiro'):
+            a, u, s, e = process_multi(parser, paths, cli)
+        else:
+            a, u, s, e = process(parser, paths)
+        
+        log.info(f"[{cli.capitalize()}] 新增:{a} 更新:{u} 跳过:{s} 错误:{e}")
+        update_cli_meta(cli)
+        total_added += a
+        total_updated += u
+        total_skipped += s
+        total_errors += e
+    
+    total = storage.get_total_stats()
+    storage.update_total_meta(total['sessions'], total['messages'], total['tokens'])
+    
+    log.info("=" * 50)
+    log.info(f"总计: {total['sessions']} 会话, {total['messages']} 消息")
+    if total_errors > 0:
+        log.warning(f"错误: {total_errors} 个文件解析失败")
+    log.info("✓ 同步完成!")
+    
+    print_token_stats()
+
+def cmd_watch():
+    from watcher import ChatWatcher
+    from datetime import datetime
+    
+    log = get_logger()
+    log.info("")
+    log.info("=" * 50)
+    log.info("实时监听模式 (watchdog)")
+    log.info("=" * 50)
+    
+    watch_paths = []
+    path_source_map = {}
+    
+    for cli, key in [('codex', 'codex_paths'), ('kiro', 'kiro_paths'), 
+                     ('gemini', 'gemini_paths'), ('claude', 'claude_paths')]:
+        for p in CONFIG.get(key, []):
+            if os.path.isdir(p) or os.path.isfile(p):
+                watch_paths.append(p)
+                path_source_map[p] = cli
+    
+    def on_change(file_path, event_type):
+        now = datetime.now().strftime('%H:%M:%S')
+        source = None
+        for p, s in path_source_map.items():
+            if file_path.startswith(p) or file_path == p:
+                source = s
+                break
+        if not source:
+            return
+        
+        try:
+            if source == 'kiro':
+                parser = KiroParser()
+                for sess in parser.parse_file(file_path):
+                    storage.upsert_session(sess.session_id, sess.source, sess.file_path, sess.cwd, sess.messages, int(sess.file_mtime))
+                log.info(f"[{now}] kiro 更新")
+            elif source == 'claude':
+                parser = ClaudeParser()
+                for sess in parser.parse_file(file_path):
+                    fp = f"claude:{sess.session_id}"
+                    storage.upsert_session(sess.session_id, sess.source, fp, sess.cwd, sess.messages, int(sess.file_mtime))
+                log.info(f"[{now}] claude 更新")
+            else:
+                parser = CodexParser(source) if source == 'codex' else GeminiParser()
+                sess = parser.parse_file(file_path)
+                fp = os.path.abspath(sess.file_path)
+                storage.upsert_session(sess.session_id, sess.source, fp, sess.cwd, sess.messages, int(sess.file_mtime))
+                log.info(f"[{now}] {source} {event_type}: {os.path.basename(file_path)}")
+            
+            update_cli_meta(source)
+            total = storage.get_total_stats()
+            storage.update_total_meta(total['sessions'], total['messages'], total['tokens'])
+        except Exception as e:
+            log.error(f"[{now}] 处理失败 {file_path}: {e}")
+    
+    log.info(f"监听目录: {len(watch_paths)} 个")
+    watcher = ChatWatcher(watch_paths, on_change)
+    watcher.start()
+
+def cmd_prune():
+    log = get_logger()
+    log.info("清理孤立记录...")
+    removed = storage.prune()
+    total = sum(removed.values())
+    if total > 0:
+        for cli, count in removed.items():
+            if count > 0:
+                log.info(f"  {cli}: 删除 {count} 条")
+        log.info(f"✓ 共清理 {total} 条孤立记录")
+    else:
+        log.info("✓ 无孤立记录")
+
+def cmd_stats():
+    log = get_logger()
+    meta = storage.get_total_meta()
+    tokens = storage.get_token_stats()
+    
+    log.info("=" * 50)
+    log.info("统计信息")
+    log.info("=" * 50)
+    log.info(f"数据库: {CONFIG['db_path']}")
+    log.info(f"总会话: {meta['total_sessions']}")
+    log.info(f"总消息: {meta['total_messages']}")
+    log.info(f"最后同步: {meta['last_sync']}")
+    log.info("")
+    log.info("Token 统计 (tiktoken):")
+    total_tokens = 0
+    for source in ['codex', 'kiro', 'gemini', 'claude']:
+        t = tokens.get(source, 0)
+        if t > 0:
+            log.info(f"  {source}: {t:,}")
+            total_tokens += t
+    log.info(f"  总计: {total_tokens:,}")
+
+def cmd_search(keyword: str, source: str = None):
+    log = get_logger()
+    results = storage.search(keyword, source)
+    log.info(f"搜索 '{keyword}' 找到 {len(results)} 个会话:")
+    for r in results[:20]:
+        log.info(f"  [{r['source']}] {r['session_id']} - {r['cwd'] or 'N/A'}")
+
+def cmd_export(fmt: str, source: str = None, output: str = None):
+    log = get_logger()
+    if not output:
+        output = os.path.join(CONFIG["output_dir"], f"export.{fmt}")
+    
+    if fmt == 'json':
+        count = storage.export_json(output, source)
+    else:
+        count = storage.export_csv(output, source)
+    
+    log.info(f"✓ 导出 {count} 条到 {output}")
+
+def print_token_stats():
+    log = get_logger()
+    tokens = storage.get_token_stats()
+    log.info("")
+    log.info("=== Token 统计 (tiktoken) ===")
+    total = 0
+    for source in ['codex', 'kiro', 'gemini', 'claude']:
+        t = tokens.get(source, 0)
+        if t > 0:
+            log.info(f"  {source}: {t:,} tokens")
+            total += t
+    log.info(f"  总计: {total:,} tokens")
+
+def update_cli_meta(cli: str):
+    stats = storage.get_cli_stats(cli)
+    path = CONFIG.get(f"{cli}_paths", [""])[0] if CONFIG.get(f"{cli}_paths") else ""
+    storage.update_cli_meta(cli, path, stats['sessions'], stats['messages'], stats['tokens'])
+
+def process(parser, paths) -> tuple:
+    log = get_logger()
+    added, updated, skipped, errors = 0, 0, 0, 0
+    for f in parser.find_files(paths):
+        try:
+            s = parser.parse_file(f)
+            file_path = os.path.abspath(s.file_path)
+            db_mtime = storage.get_file_mtime(file_path)
+            file_mtime = int(s.file_mtime)
+            if db_mtime == 0:
+                storage.upsert_session(s.session_id, s.source, file_path, s.cwd, s.messages, file_mtime)
+                added += 1
+            elif file_mtime > db_mtime:
+                storage.upsert_session(s.session_id, s.source, file_path, s.cwd, s.messages, file_mtime)
+                updated += 1
+            else:
+                skipped += 1
+        except Exception as e:
+            log.debug(f"解析失败 {f}: {e}")
+            errors += 1
+    return added, updated, skipped, errors
+
+def process_multi(parser, paths, source: str) -> tuple:
+    """处理返回多个会话的解析器（Claude/Kiro）"""
+    log = get_logger()
+    added, updated, skipped, errors = 0, 0, 0, 0
+    for f in parser.find_files(paths):
+        try:
+            for s in parser.parse_file(f):
+                file_path = s.file_path  # kiro:xxx 或 claude:xxx
+                db_mtime = storage.get_file_mtime(file_path)
+                file_mtime = int(s.file_mtime)
+                if db_mtime == 0:
+                    storage.upsert_session(s.session_id, s.source, file_path, s.cwd, s.messages, file_mtime)
+                    added += 1
+                elif file_mtime > db_mtime:
+                    storage.upsert_session(s.session_id, s.source, file_path, s.cwd, s.messages, file_mtime)
+                    updated += 1
+                else:
+                    skipped += 1
+        except Exception as e:
+            log.debug(f"解析失败 {f}: {e}")
+            errors += 1
+    return added, updated, skipped, errors
+
+if __name__ == '__main__':
+    main()
--- a/libs/external/chat-vault/src/parsers/init.py
+++ b/libs/external/chat-vault/src/parsers/init.py
@ -0,0 +1,7 @@
+from .codex import CodexParser
+from .gemini import GeminiParser
+from .claude import ClaudeParser
+from .kiro import KiroParser
+from .base import SessionData
+
+__all__ = ["CodexParser", "GeminiParser", "ClaudeParser", "KiroParser", "SessionData"]
--- a/libs/external/chat-vault/src/parsers/base.py
+++ b/libs/external/chat-vault/src/parsers/base.py
@ -0,0 +1,24 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+from abc import ABC, abstractmethod
+from dataclasses import dataclass, field
+from typing import List, Dict
+
+@dataclass
+class SessionData:
+    """会话数据"""
+    session_id: str
+    source: str
+    file_path: str
+    file_mtime: float = 0
+    cwd: str = None
+    messages: List[Dict] = field(default_factory=list)  # [{"time", "role", "content"}]
+
+class BaseParser(ABC):
+    @abstractmethod
+    def find_files(self, paths: list) -> list:
+        pass
+    
+    @abstractmethod
+    def parse_file(self, filepath: str) -> SessionData:
+        pass
--- a/libs/external/chat-vault/src/parsers/claude.py
+++ b/libs/external/chat-vault/src/parsers/claude.py
@ -0,0 +1,54 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+import os
+import json
+import hashlib
+from datetime import datetime
+from collections import defaultdict
+from .base import BaseParser, SessionData
+
+class ClaudeParser(BaseParser):
+    def find_files(self, paths: list) -> list:
+        files = []
+        for base in paths:
+            history = os.path.join(base, "history.jsonl")
+            if os.path.exists(history):
+                files.append(history)
+        return files
+    
+    def parse_file(self, filepath: str) -> list:
+        """返回多个 SessionData（按 project 分组）"""
+        projects = defaultdict(list)
+        file_mtime = os.path.getmtime(filepath)
+        
+        with open(filepath, 'r', encoding='utf-8') as f:
+            for line in f:
+                line = line.strip()
+                if not line:
+                    continue
+                data = json.loads(line)
+                content = data.get('display', '')
+                if not content:
+                    continue
+                
+                project = data.get('project', 'unknown')
+                ts_ms = data.get('timestamp', 0)
+                ts = datetime.fromtimestamp(ts_ms / 1000).isoformat() if ts_ms else ''
+                
+                projects[project].append({
+                    'time': ts,
+                    'role': 'user',
+                    'content': content
+                })
+        
+        return [
+            SessionData(
+                session_id='claude-' + hashlib.md5(proj.encode()).hexdigest()[:12],
+                source='claude',
+                file_path=filepath,
+                file_mtime=file_mtime,
+                cwd=proj,
+                messages=msgs
+            )
+            for proj, msgs in projects.items()
+        ]
--- a/libs/external/chat-vault/src/parsers/codex.py
+++ b/libs/external/chat-vault/src/parsers/codex.py
@ -0,0 +1,70 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+import os
+import json
+import re
+from .base import BaseParser, SessionData
+
+class CodexParser(BaseParser):
+    def __init__(self, source: str = 'codex'):
+        self.source = source
+    
+    def find_files(self, paths: list) -> list:
+        files = []
+        for base in paths:
+            if not os.path.exists(base):
+                continue
+            for root, _, names in os.walk(base):
+                for f in names:
+                    if f.endswith('.jsonl') and f != 'history.jsonl':
+                        files.append(os.path.join(root, f))
+        return files
+    
+    def _extract_id(self, filepath: str) -> str:
+        match = re.search(r'([0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12})', 
+                          os.path.basename(filepath))
+        return match.group(1) if match else os.path.basename(filepath).replace('.jsonl', '')
+    
+    def parse_file(self, filepath: str) -> SessionData:
+        s = SessionData(
+            session_id=self._extract_id(filepath),
+            source=self.source,
+            file_path=filepath,
+            file_mtime=os.path.getmtime(filepath)
+        )
+        
+        with open(filepath, 'r', encoding='utf-8', errors='ignore') as f:
+            for line in f:
+                line = line.strip()
+                if not line or line[0] != '{':
+                    continue
+                try:
+                    data = json.loads(line)
+                except json.JSONDecodeError:
+                    continue
+                
+                if data.get('type') == 'session_meta':
+                    p = data.get('payload', {})
+                    s.cwd = p.get('cwd')
+                    s.session_id = p.get('id', s.session_id)
+                    continue
+                
+                if data.get('type') != 'response_item':
+                    continue
+                payload = data.get('payload', {})
+                if payload.get('type') != 'message':
+                    continue
+                role = payload.get('role')
+                if role not in ('user', 'assistant'):
+                    continue
+                
+                parts = [item.get('text', '') for item in payload.get('content', [])
+                         if isinstance(item, dict) and item.get('type') in ('input_text', 'output_text', 'text')]
+                if parts:
+                    s.messages.append({
+                        'time': data.get('timestamp', ''),
+                        'role': 'user' if role == 'user' else 'ai',
+                        'content': ' '.join(parts)
+                    })
+        
+        return s
--- a/libs/external/chat-vault/src/parsers/gemini.py
+++ b/libs/external/chat-vault/src/parsers/gemini.py
@ -0,0 +1,40 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+import os
+import glob
+import json
+from .base import BaseParser, SessionData
+
+class GeminiParser(BaseParser):
+    def find_files(self, paths: list) -> list:
+        files = []
+        for base in paths:
+            if os.path.exists(base):
+                files.extend(glob.glob(os.path.join(base, "*", "chats", "*.json")))
+        return files
+    
+    def parse_file(self, filepath: str) -> SessionData:
+        s = SessionData(
+            session_id=os.path.basename(filepath).replace('.json', ''),
+            source='gemini',
+            file_path=filepath,
+            file_mtime=os.path.getmtime(filepath)
+        )
+        
+        with open(filepath, 'r', encoding='utf-8') as f:
+            data = json.load(f)
+        
+        s.session_id = data.get('sessionId', s.session_id)
+        
+        for msg in data.get('messages', []):
+            if msg.get('type') not in ('user', 'gemini'):
+                continue
+            content = msg.get('content', '')
+            if content:
+                s.messages.append({
+                    'time': msg.get('timestamp', ''),
+                    'role': 'user' if msg.get('type') == 'user' else 'ai',
+                    'content': content
+                })
+        
+        return s
--- a/libs/external/chat-vault/src/parsers/kiro.py
+++ b/libs/external/chat-vault/src/parsers/kiro.py
@ -0,0 +1,75 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""Kiro CLI 解析器 - 从 SQLite 数据库读取"""
+import os
+import json
+import sqlite3
+import hashlib
+from datetime import datetime
+from .base import BaseParser, SessionData
+
+KIRO_DB = os.path.expanduser("~/.local/share/kiro-cli/data.sqlite3")
+
+class KiroParser(BaseParser):
+    def find_files(self, paths: list) -> list:
+        """返回数据库路径（如果存在）"""
+        if os.path.exists(KIRO_DB):
+            return [KIRO_DB]
+        return []
+    
+    def parse_file(self, filepath: str) -> list:
+        """解析 Kiro SQLite 数据库，返回多个 SessionData"""
+        sessions = []
+        file_mtime = os.path.getmtime(filepath)
+        
+        conn = sqlite3.connect(filepath)
+        for row in conn.execute('SELECT key, value FROM conversations'):
+            cwd, value = row
+            try:
+                data = json.loads(value)
+            except json.JSONDecodeError:
+                continue
+            
+            conv_id = data.get('conversation_id', hashlib.md5(cwd.encode()).hexdigest()[:12])
+            history = data.get('history', [])
+            
+            messages = []
+            for item in history:
+                # 用户消息
+                if 'user' in item:
+                    user = item['user']
+                    content = user.get('content', {})
+                    if isinstance(content, dict) and 'Prompt' in content:
+                        prompt = content['Prompt'].get('prompt', '')
+                        if prompt:
+                            messages.append({
+                                'time': '',
+                                'role': 'user',
+                                'content': prompt
+                            })
+                
+                # AI 回复
+                if 'assistant' in item:
+                    assistant = item['assistant']
+                    content = assistant.get('content', {})
+                    if isinstance(content, dict) and 'Message' in content:
+                        msg = content['Message'].get('message', '')
+                        if msg:
+                            messages.append({
+                                'time': '',
+                                'role': 'ai',
+                                'content': msg
+                            })
+            
+            if messages:
+                sessions.append(SessionData(
+                    session_id=f'kiro-{conv_id[:12]}',
+                    source='kiro',
+                    file_path=f'kiro:{conv_id}',
+                    file_mtime=file_mtime,
+                    cwd=cwd,
+                    messages=messages
+                ))
+        
+        conn.close()
+        return sessions
--- a/libs/external/chat-vault/src/storage.py
+++ b/libs/external/chat-vault/src/storage.py
@ -0,0 +1,246 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""SQLite 存储模块 - 完整版"""
+import sqlite3
+import json
+import os
+import datetime
+import tiktoken
+
+SCHEMA_VERSION = 5
+CLIS = ('codex', 'kiro', 'gemini', 'claude')
+
+_encoder = tiktoken.get_encoding("cl100k_base")
+
+def count_tokens(text: str) -> int:
+    return len(_encoder.encode(text)) if text else 0
+
+class ChatStorage:
+    def __init__(self, db_path: str):
+        self.db_path = db_path
+        os.makedirs(os.path.dirname(db_path) or '.', exist_ok=True)
+        self._init_db()
+    
+    def _conn(self):
+        return sqlite3.connect(self.db_path)
+    
+    def _init_db(self):
+        with self._conn() as conn:
+            conn.execute('''CREATE TABLE IF NOT EXISTS meta (key TEXT PRIMARY KEY, value TEXT)''')
+            for cli in CLIS:
+                conn.execute(f'''CREATE TABLE IF NOT EXISTS meta_{cli} (key TEXT PRIMARY KEY, value TEXT)''')
+            conn.execute('''
+                CREATE TABLE IF NOT EXISTS sessions (
+                    file_path TEXT PRIMARY KEY,
+                    session_id TEXT,
+                    source TEXT NOT NULL,
+                    cwd TEXT,
+                    messages TEXT,
+                    file_mtime INTEGER,
+                    start_time TEXT,
+                    token_count INTEGER DEFAULT 0
+                )
+            ''')
+            conn.execute('CREATE INDEX IF NOT EXISTS idx_source ON sessions(source)')
+            conn.execute('CREATE INDEX IF NOT EXISTS idx_session_id ON sessions(session_id)')
+            conn.execute('CREATE INDEX IF NOT EXISTS idx_start_time ON sessions(start_time)')
+            self._set_meta('meta', 'schema_version', str(SCHEMA_VERSION))
+    
+    def _set_meta(self, table: str, key: str, value: str):
+        with self._conn() as conn:
+            conn.execute(f'INSERT OR REPLACE INTO {table} (key, value) VALUES (?, ?)', (key, value))
+    
+    def _get_meta(self, table: str, key: str) -> str:
+        with self._conn() as conn:
+            row = conn.execute(f'SELECT value FROM {table} WHERE key = ?', (key,)).fetchone()
+            return row[0] if row else None
+    
+    def update_cli_meta(self, cli: str, path: str, sessions: int, messages: int, tokens: int = None):
+        table = f'meta_{cli}'
+        now = datetime.datetime.now().isoformat()
+        # 顺序: path, sessions, messages, total_tokens, last_sync
+        self._set_meta(table, 'path', path)
+        self._set_meta(table, 'sessions', str(sessions))
+        self._set_meta(table, 'messages', str(messages))
+        self._set_meta(table, 'total_tokens', str(tokens or 0))
+        self._set_meta(table, 'last_sync', now)
+    
+    def update_total_meta(self, sessions: int, messages: int, tokens: int = None):
+        now = datetime.datetime.now().isoformat()
+        self._set_meta('meta', 'total_sessions', str(sessions))
+        self._set_meta('meta', 'total_messages', str(messages))
+        if tokens is not None:
+            self._set_meta('meta', 'total_tokens', str(tokens))
+        self._set_meta('meta', 'last_sync', now)
+    
+    def get_total_meta(self) -> dict:
+        return {
+            'schema_version': int(self._get_meta('meta', 'schema_version') or 0),
+            'total_sessions': int(self._get_meta('meta', 'total_sessions') or 0),
+            'total_messages': int(self._get_meta('meta', 'total_messages') or 0),
+            'last_sync': self._get_meta('meta', 'last_sync'),
+        }
+    
+    def get_file_mtime(self, file_path: str) -> int:
+        with self._conn() as conn:
+            row = conn.execute('SELECT file_mtime FROM sessions WHERE file_path = ?', (file_path,)).fetchone()
+            return row[0] if row else 0
+    
+    def upsert_session(self, session_id: str, source: str, file_path: str,
+                       cwd: str, messages: list, file_mtime: int, start_time: str = None):
+        if file_path and not file_path.startswith('claude:') and not os.path.isabs(file_path):
+            file_path = os.path.abspath(file_path)
+        
+        total_tokens = sum(count_tokens(msg.get('content', '')) for msg in messages)
+        if not start_time and messages:
+            start_time = messages[0].get('time')
+        
+        messages_json = json.dumps(messages, ensure_ascii=False)
+        with self._conn() as conn:
+            conn.execute('''
+                INSERT INTO sessions (file_path, session_id, source, cwd, messages, file_mtime, start_time, token_count)
+                VALUES (?, ?, ?, ?, ?, ?, ?, ?)
+                ON CONFLICT(file_path) DO UPDATE SET
+                    session_id=excluded.session_id, messages=excluded.messages,
+                    file_mtime=excluded.file_mtime, start_time=excluded.start_time, token_count=excluded.token_count
+            ''', (file_path, session_id, source, cwd, messages_json, file_mtime, start_time, total_tokens))
+    
+    def get_cli_stats(self, cli: str) -> dict:
+        with self._conn() as conn:
+            sessions = conn.execute('SELECT COUNT(*) FROM sessions WHERE source = ?', (cli,)).fetchone()[0]
+            row = conn.execute('SELECT SUM(json_array_length(messages)) FROM sessions WHERE source = ?', (cli,)).fetchone()
+            messages = row[0] or 0
+            tokens = conn.execute('SELECT SUM(token_count) FROM sessions WHERE source = ?', (cli,)).fetchone()[0] or 0
+        return {'sessions': sessions, 'messages': messages, 'tokens': tokens}
+    
+    def get_total_stats(self) -> dict:
+        with self._conn() as conn:
+            sessions = conn.execute('SELECT COUNT(*) FROM sessions').fetchone()[0]
+            row = conn.execute('SELECT SUM(json_array_length(messages)) FROM sessions').fetchone()
+            messages = row[0] or 0
+            tokens = conn.execute('SELECT SUM(token_count) FROM sessions').fetchone()[0] or 0
+        return {'sessions': sessions, 'messages': messages, 'tokens': tokens}
+    
+    def get_token_stats(self) -> dict:
+        with self._conn() as conn:
+            rows = conn.execute('SELECT source, SUM(token_count) FROM sessions GROUP BY source').fetchall()
+        return {r[0]: r[1] or 0 for r in rows}
+    
+    # === 清理孤立记录 ===
+    def prune(self) -> dict:
+        """删除源文件已不存在的记录"""
+        removed = {'codex': 0, 'kiro': 0, 'gemini': 0, 'claude': 0}
+        with self._conn() as conn:
+            rows = conn.execute('SELECT file_path, source FROM sessions').fetchall()
+            for fp, source in rows:
+                if fp.startswith('claude:'):
+                    continue  # Claude 使用虚拟路径
+                if not os.path.exists(fp):
+                    conn.execute('DELETE FROM sessions WHERE file_path = ?', (fp,))
+                    removed[source] = removed.get(source, 0) + 1
+        return removed
+    
+    # === 查询 ===
+    def search(self, keyword: str, source: str = None, limit: int = 50) -> list:
+        """搜索消息内容"""
+        sql = "SELECT file_path, session_id, source, cwd, messages, start_time FROM sessions WHERE messages LIKE ?"
+        params = [f'%{keyword}%']
+        if source:
+            sql += " AND source = ?"
+            params.append(source)
+        sql += f" ORDER BY start_time DESC LIMIT {limit}"
+        
+        results = []
+        with self._conn() as conn:
+            for row in conn.execute(sql, params):
+                results.append({
+                    'file_path': row[0], 'session_id': row[1], 'source': row[2],
+                    'cwd': row[3], 'messages': json.loads(row[4]), 'start_time': row[5]
+                })
+        return results
+    
+    def get_session(self, file_path: str) -> dict:
+        """获取单个会话"""
+        with self._conn() as conn:
+            row = conn.execute(
+                'SELECT file_path, session_id, source, cwd, messages, start_time, token_count FROM sessions WHERE file_path = ?',
+                (file_path,)
+            ).fetchone()
+        if not row:
+            return None
+        return {
+            'file_path': row[0], 'session_id': row[1], 'source': row[2], 'cwd': row[3],
+            'messages': json.loads(row[4]), 'start_time': row[5], 'token_count': row[6]
+        }
+    
+    def list_sessions(self, source: str = None, limit: int = 100, offset: int = 0) -> list:
+        """列出会话"""
+        sql = "SELECT file_path, session_id, source, cwd, start_time, token_count FROM sessions"
+        params = []
+        if source:
+            sql += " WHERE source = ?"
+            params.append(source)
+        sql += f" ORDER BY start_time DESC LIMIT {limit} OFFSET {offset}"
+        
+        results = []
+        with self._conn() as conn:
+            for row in conn.execute(sql, params):
+                results.append({
+                    'file_path': row[0], 'session_id': row[1], 'source': row[2],
+                    'cwd': row[3], 'start_time': row[4], 'token_count': row[5]
+                })
+        return results
+    
+    # === 导出 ===
+    def export_json(self, output_path: str, source: str = None):
+        """导出为 JSON"""
+        sql = "SELECT file_path, session_id, source, cwd, messages, start_time, token_count FROM sessions"
+        params = []
+        if source:
+            sql += " WHERE source = ?"
+            params.append(source)
+        sql += " ORDER BY start_time"
+        
+        data = []
+        with self._conn() as conn:
+            for row in conn.execute(sql, params):
+                data.append({
+                    'file_path': row[0], 'session_id': row[1], 'source': row[2], 'cwd': row[3],
+                    'messages': json.loads(row[4]), 'start_time': row[5], 'token_count': row[6]
+                })
+        
+        with open(output_path, 'w', encoding='utf-8') as f:
+            json.dump(data, f, ensure_ascii=False, indent=2)
+        return len(data)
+    
+    def export_csv(self, output_path: str, source: str = None):
+        """导出为 CSV（扁平化消息）"""
+        import csv
+        sql = "SELECT session_id, source, cwd, messages, start_time FROM sessions"
+        params = []
+        if source:
+            sql += " WHERE source = ?"
+            params.append(source)
+        sql += " ORDER BY start_time"
+        
+        count = 0
+        with open(output_path, 'w', encoding='utf-8', newline='') as f:
+            writer = csv.writer(f)
+            writer.writerow(['session_id', 'source', 'cwd', 'time', 'role', 'content'])
+            with self._conn() as conn:
+                for row in conn.execute(sql, params):
+                    session_id, src, cwd, msgs_json, _ = row
+                    for msg in json.loads(msgs_json):
+                        writer.writerow([session_id, src, cwd, msg.get('time', ''), msg.get('role', ''), msg.get('content', '')])
+                        count += 1
+        return count
+    
+    # === 获取所有文件路径（用于 prune 检查） ===
+    def get_all_file_paths(self, source: str = None) -> set:
+        sql = "SELECT file_path FROM sessions"
+        params = []
+        if source:
+            sql += " WHERE source = ?"
+            params.append(source)
+        with self._conn() as conn:
+            return {row[0] for row in conn.execute(sql, params)}
--- a/libs/external/chat-vault/src/watcher.py
+++ b/libs/external/chat-vault/src/watcher.py
@ -0,0 +1,44 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""跨平台文件监控 (Linux/macOS/Windows)"""
+
+from watchdog.observers import Observer
+from watchdog.events import FileSystemEventHandler
+import time
+
+class ChatFileHandler(FileSystemEventHandler):
+    def __init__(self, callback, extensions):
+        self.callback = callback
+        self.extensions = extensions
+    
+    def _check(self, event):
+        if event.is_directory:
+            return
+        path = event.src_path
+        if any(path.endswith(ext) for ext in self.extensions):
+            self.callback(path, event.event_type)
+    
+    def on_created(self, event):
+        self._check(event)
+    
+    def on_modified(self, event):
+        self._check(event)
+
+class ChatWatcher:
+    def __init__(self, paths: list, callback, extensions=('.jsonl', '.json')):
+        self.observer = Observer()
+        handler = ChatFileHandler(callback, extensions)
+        for path in paths:
+            self.observer.schedule(handler, path, recursive=True)
+    
+    def start(self):
+        self.observer.start()
+        try:
+            while True:
+                time.sleep(1)
+        except KeyboardInterrupt:
+            self.stop()
+    
+    def stop(self):
+        self.observer.stop()
+        self.observer.join()
--- a/libs/external/chat-vault/start.bat
+++ b/libs/external/chat-vault/start.bat
@ -0,0 +1,7 @@
+@echo off
+cd /d "%~dp0src"
+echo 正在启动 AI Chat Converter...
+echo 按 Ctrl+C 停止
+echo.
+python main.py --watch
+pause
--- a/libs/external/chat-vault/start.sh
+++ b/libs/external/chat-vault/start.sh
@ -0,0 +1,7 @@
+#!/bin/bash
+# AI Chat Converter - 一键启动
+cd "$(dirname "$0")/src"
+echo "正在启动 AI Chat Converter..."
+echo "按 Ctrl+C 停止"
+echo ""
+python3 main.py --watch
--- a/scripts/.gitkeep
+++ b/scripts/.gitkeep