debug-mcp-server

debug-mcp-server

Bridges web browsers and Claude Code for seamless debugging workflows by receiving debug data from browser widgets and exposing it to Claude Code via MCP protocol.

Category
访问服务器

README

Debug MCP Server

An MCP (Model Context Protocol) server that bridges web browsers and Claude Code for seamless debugging workflows.

Overview

This server receives debug data from browser widgets (console logs, error stacks, screenshots) and exposes it to Claude Code via the MCP protocol. It also facilitates interactive Q&A between Claude Code and browser users.

Features

  • Debug Data Ingestion: HTTP endpoints to receive debug reports from browsers
  • MCP Tools: Expose debug data to Claude Code through MCP protocol
  • Interactive Q&A: WebSocket-based bidirectional communication for clarification questions
  • File Storage: Organized filesystem storage for reports and screenshots
  • Localhost Only: Secure local-only server for development use

Quick Start

Installation

# Global installation
npm install -g @yourco/debug-mcp-server

# Or use npx
npx @yourco/debug-mcp-server

Running the Server

# Start the server
debug-mcp-server

# Or with custom port
DEBUG_MCP_PORT=4000 debug-mcp-server

Configuration

Create ~/.debug-mcp/config.json:

{
  "port": 3000,
  "storagePath": "~/.debug-mcp",
  "retentionDays": 7
}

Or use environment variables:

  • DEBUG_MCP_PORT - HTTP server port (default: 3000)
  • DEBUG_MCP_STORAGE_PATH - Storage directory (default: ~/.debug-mcp)
  • DEBUG_MCP_RETENTION_DAYS - Report retention period (default: 7)

MCP Integration

Connect Claude Code to Server

Add to your Claude Code MCP configuration:

{
  "mcpServers": {
    "debug-reports": {
      "command": "node",
      "args": ["/path/to/debug-mcp-server/dist/index.js"]
    }
  }
}

Available MCP Tools

  • submit_debug_report - Receive debug data from browser
  • list_debug_reports - List all stored debug reports
  • get_debug_report - Retrieve specific debug report by ID
  • ask_user_questions - Send questions to browser and wait for answers
  • clear_old_reports - Clean up old debug data

Available MCP Resources

  • debug://reports/latest - Get most recent debug report
  • debug://reports/{id} - Get specific report by ID
  • debug://screenshots/{id} - Access screenshot files

API Endpoints

POST /api/debug

Receive debug data from browser widget.

Request Body:

{
  "logs": [
    { "level": "error", "message": "Error message", "timestamp": 1234567890 }
  ],
  "error": {
    "message": "TypeError: Cannot read property...",
    "stack": "Error stack trace..."
  },
  "screenshot": "data:image/png;base64,...",
  "comment": "User description of the issue",
  "url": "https://example.com/page",
  "timestamp": 1234567890
}

Response:

{
  "success": true,
  "data": {
    "reportId": "abc123",
    "timestamp": 1234567890
  }
}

POST /api/questions/answer

Receive answers from browser to Claude's questions.

Request Body:

{
  "sessionId": "session-123",
  "answers": [
    { "questionId": "q1", "answer": "User's answer" }
  ]
}

Development

Setup

# Clone the repository
git clone https://github.com/yourco/debug-mcp-server.git
cd debug-mcp-server

# Install dependencies
npm install

# Run in development mode
npm run dev

Project Structure

debug-mcp-server/
├── src/
│   ├── index.ts              # Entry point
│   ├── mcp/                  # MCP server implementation
│   ├── http/                 # HTTP/WebSocket servers
│   ├── storage/              # File storage layer
│   ├── types/                # TypeScript types
│   └── config/               # Configuration
├── specs/                    # SpecKit specifications
├── tests/                    # Jest tests
└── logs/                     # Session logs

Testing

# Run unit tests
npm test

# Run with coverage
npm run test:coverage

# Run linter
npm run lint

Building

# Build for production
npm run build

# Output: dist/index.js

Architecture

┌─────────────────────┐
│  Browser Widget     │
│  (HTTP/WebSocket)   │
└──────────┬──────────┘
           │
           ↓
┌─────────────────────────────────────┐
│  Debug MCP Server                   │
│  ┌─────────────────────────────┐    │
│  │ HTTP Server (Express)       │    │
│  │  - POST /api/debug          │    │
│  │  - POST /api/questions      │    │
│  └─────────────────────────────┘    │
│  ┌─────────────────────────────┐    │
│  │ WebSocket Server            │    │
│  │  - Push questions to browser│    │
│  └─────────────────────────────┘    │
│  ┌─────────────────────────────┐    │
│  │ MCP Server                  │    │
│  │  - Tools & Resources        │    │
│  └─────────────────────────────┘    │
│  ┌─────────────────────────────┐    │
│  │ File Storage                │    │
│  │  - ~/.debug-mcp/            │    │
│  └─────────────────────────────┘    │
└──────────┬──────────────────────────┘
           │
           ↓
┌─────────────────────┐
│  Claude Code        │
│  (VS Code)          │
└─────────────────────┘

Storage Structure

~/.debug-mcp/
├── reports/
│   ├── 2025-11-26/
│   │   ├── report-abc123.json
│   │   └── report-def456.json
│   └── 2025-11-27/
│       └── report-ghi789.json
├── screenshots/
│   ├── screenshot-abc123.png
│   └── screenshot-def456.png
└── config.json

Troubleshooting

Server won't start

Check if port is already in use:

lsof -i :3000

Use a different port:

DEBUG_MCP_PORT=4000 debug-mcp-server

Claude Code can't connect

  1. Verify server is running: curl http://localhost:3000/api/health
  2. Check MCP configuration in Claude Code settings
  3. Restart VS Code after config changes

Browser widget can't send data

  1. Verify server is running on localhost:3000
  2. Check browser console for CORS errors
  3. Ensure browser has network access to localhost

Contributing

See CLAUDE.md for development guidelines and project constitution.

License

MIT

Support

For issues and feature requests, please open an issue on GitHub.

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选