mcp-dev-tools
Aggregates GitHub, GitLab, and Figma development tools via MCP, enabling file access, code search, and design inspection. Also supports executing standalone tasks with Claude Code or Codex CLI.
README
mcp-dev-tools
通过 MCP (Model Context Protocol) 聚合 GitHub / GitLab / Figma 等开发工具的 Server。
支持所有兼容 MCP 的客户端:Claude Code、Codex、VS Code (Copilot)、Kiro、Cursor 等。
提供的工具
| 工具 | 说明 |
|---|---|
github_get_file |
获取 GitHub 仓库中指定文件的内容 |
github_list_files |
列出 GitHub 仓库的文件目录 |
github_search_code |
搜索 GitHub 仓库代码 |
gitlab_get_file |
获取 GitLab 仓库中指定文件的内容 |
gitlab_list_files |
列出 GitLab 仓库的文件目录 |
gitlab_search_code |
搜索 GitLab 仓库代码 |
figma_inspect |
按前缀查找 Figma 设计稿中的节点 |
figma_tree |
浏览 Figma 节点树(支持过滤) |
figma_images |
获取 Figma 节点导出图片 URL |
figma_export |
批量导出 Figma 节点图片(返回 URL + 映射) |
figma_text |
提取 Figma 设计稿中的文字、颜色和高亮 |
codex_exec |
调用 Codex CLI 执行独立任务 |
claude_exec |
调用 Claude Code CLI 执行独立任务,支持 direct / relay 模式 |
快速安装
需要 Node.js ≥ 18.0,可通过
node -v检查版本。
git clone https://github.com/dan-0526/mcp-dev-tools.git ~/mcp-dev-tools
cd ~/mcp-dev-tools
npm install
获取 Token
按需准备,不用的平台留空即可:
- GitHub: https://github.com/settings/tokens → Generate new token (classic) → 勾选
repo - GitLab: GitLab → Preferences → Access Tokens → 勾选
read_repository - Figma: Figma → 左上角头像 → Settings → Personal access tokens → Generate new token
⚠️ Token 不要提交到仓库,通过各客户端的 env 配置传入。
Claude Relay
claude_exec 默认使用 direct 模式:MCP server 进程直接执行 claude -p。默认会清理 HTTP_PROXY / HTTPS_PROXY / http_proxy / https_proxy,避免本机代理环境影响 Claude Code CLI。确实需要保留代理时,可以传 preserveProxyEnv: true。
如果 MCP 客户端所在环境无法直接运行 Claude,可以改用 relay 模式:在正常网络环境中单独启动 relay server,让 MCP 通过 HTTP 转发任务。
启动 relay:
cd /path/to/mcp-dev-tools
CLAUDE_RELAY_TOKEN=<shared-secret> npm run claude-relay
默认监听 http://127.0.0.1:38765。如果客户端不能访问 127.0.0.1,可以显式设置:
CLAUDE_RELAY_HOST=<reachable-host>
CLAUDE_RELAY_PORT=38765
CLAUDE_RELAY_TOKEN=<shared-secret>
npm run claude-relay
MCP server 侧配置:
CLAUDE_EXEC_MODE=relay
CLAUDE_RELAY_URL=http://<reachable-host>:38765
CLAUDE_RELAY_TOKEN=<shared-secret>
安全提醒:不要在无 token 的情况下把 relay 绑定到局域网地址或 0.0.0.0。
客户端配置
选择你使用的客户端,复制命令到终端执行。执行前先替换 <...> 占位符为你自己的值。
路径示例:如果你 clone 到了
~/mcp-dev-tools,那绝对路径就是/Users/yourname/mcp-dev-tools/src/index.js,macOS 下可以通过pwd命令查看。
Claude Code
claude mcp add dev-tools \
-e GITHUB_TOKEN=<your-github-token> \
-e GITLAB_TOKEN=<your-gitlab-token> \
-e GITLAB_URL=https://gitlab.com \
-e FIGMA_TOKEN=<your-figma-token> \
-- node <绝对路径>/mcp-dev-tools/src/index.js
VS Code (GitHub Copilot)
在项目根目录执行:
mkdir -p .vscode
cat > .vscode/mcp.json << 'EOF'
{
"servers": {
"dev-tools": {
"type": "stdio",
"command": "node",
"args": ["<绝对路径>/mcp-dev-tools/src/index.js"],
"env": {
"GITHUB_TOKEN": "<your-github-token>",
"GITLAB_TOKEN": "<your-gitlab-token>",
"GITLAB_URL": "https://gitlab.com",
"FIGMA_TOKEN": "<your-figma-token>"
}
}
}
}
EOF
Codex CLI
mkdir -p ~/.codex
cat >> ~/.codex/config.toml << 'EOF'
[mcp_servers.dev-tools]
command = "node"
args = ["<绝对路径>/mcp-dev-tools/src/index.js"]
env = { "GITHUB_TOKEN" = "<your-github-token>", "GITLAB_TOKEN" = "<your-gitlab-token>", "GITLAB_URL" = "https://gitlab.com", "FIGMA_TOKEN" = "<your-figma-token>" }
EOF
Kiro / Cursor
这两个客户端格式相同,区别只在配置文件路径:
| 客户端 | 配置文件路径 |
|---|---|
| Kiro(全局) | ~/.kiro/settings/mcp.json |
| Kiro(工作区) | .kiro/settings/mcp.json |
| Cursor | ~/.cursor/mcp.json |
配置内容(以 Kiro 全局为例):
mkdir -p ~/.kiro/settings
cat > ~/.kiro/settings/mcp.json << 'EOF'
{
"mcpServers": {
"dev-tools": {
"command": "node",
"args": ["<绝对路径>/mcp-dev-tools/src/index.js"],
"env": {
"GITHUB_TOKEN": "<your-github-token>",
"GITLAB_TOKEN": "<your-gitlab-token>",
"GITLAB_URL": "https://gitlab.com",
"FIGMA_TOKEN": "<your-figma-token>"
}
}
}
}
EOF
Cursor 只需把路径改为 ~/.cursor/mcp.json。
项目结构
mcp-dev-tools/
├── src/
│ ├── index.js # 入口,启动 MCP Server
│ ├── tools.js # 工具注册
│ ├── github.js # GitHub API 封装
│ ├── gitlab.js # GitLab API 封装
│ └── figma.js # Figma API 封装
├── .env.example
├── .gitignore
├── package.json
└── README.md
使用示例
配置好后,在聊天中直接说:
- "帮我看看 github 上 facebook/react 的 package.json"
- "列出 gitlab 项目 mygroup/myproject 的 src 目录"
- "看一下这个 Figma 设计稿的节点树:https://www.figma.com/design/xxx"
- "提取这个 Figma 页面里的所有文字内容"
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.