devscope-mcp
An MCP server that gives Claude live access to your GitHub workspace — PR reviews, issue triaging, repo search, and weekly digest reports through natural language.
README
devscope-mcp
An MCP server that gives Claude live access to your GitHub workspace — PR reviews, issue triaging, repo search, and weekly digest reports through natural language.
Built on the Model Context Protocol (MCP), devscope-mcp plugs directly into Claude Desktop (or any MCP-compatible client) and lets you work with GitHub without ever leaving the conversation.
Features
- PR Summaries — full diff metadata, comments, and review stats for any pull request
- Issue Triage — list and filter open/closed issues across any repository
- Repo Explorer — browse repos for a user or org, with stats at a glance
- Code Search — GitHub code-search with optional repo scoping
- Contributor Stats — commit volume, lines changed, and active weeks per contributor
- Weekly Digest — merged PRs, opened issues, and top contributors for the past 7 days
Tech Stack
| Layer | Technology |
|---|---|
| Protocol | Model Context Protocol (mcp >= 1.0) |
| GitHub API | PyGithub >= 2.3 |
| HTTP client | httpx >= 0.27 |
| Config | python-dotenv >= 1.0 |
| Runtime | Python 3.11+ |
| Tests | pytest 8, pytest-asyncio, pytest-mock |
| Build | Hatchling via pyproject.toml |
Setup
1. Prerequisites
- Python 3.11 or later
- A GitHub Personal Access Token with the following scopes:
repo(full repository access)read:orgread:user
2. Clone and install
git clone https://github.com/your-user/devscope-mcp.git
cd devscope-mcp
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install -e ".[dev]"
3. Configure environment
cp .env.example .env
# edit .env and set GITHUB_TOKEN (and optionally GITHUB_DEFAULT_ORG)
4. Add to Claude Desktop
Open your Claude Desktop config file:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
Add the devscope-mcp entry to mcpServers:
{
"mcpServers": {
"devscope-mcp": {
"command": "python",
"args": ["-m", "src.server"],
"cwd": "/absolute/path/to/devscope-mcp",
"env": {
"GITHUB_TOKEN": "ghp_your_token_here",
"GITHUB_DEFAULT_ORG": "your-org-name"
}
}
}
}
If you installed with pip install -e ., you can use the console script instead:
{
"mcpServers": {
"devscope-mcp": {
"command": "devscope-mcp",
"env": {
"GITHUB_TOKEN": "ghp_your_token_here",
"GITHUB_DEFAULT_ORG": "your-org-name"
}
}
}
}
Restart Claude Desktop. You should see the devscope-mcp tools available in the tools menu.
Available Tools
| Tool | Description |
|---|---|
list_repos |
List repositories for a GitHub org or authenticated user, sorted by last push |
get_repo_info |
Detailed metadata for a single repo (language, stars, forks, topics, etc.) |
summarize_pr |
Full summary of a pull request including changed files, comments, and deltas |
list_issues |
List issues with state filter (open / closed / all) and configurable limit |
search_code |
GitHub code search with optional repo scope |
get_contributor_stats |
Per-contributor commit count, lines added/deleted, and active weeks |
get_weekly_digest |
7-day activity digest: merged PRs, opened issues, top contributors |
Example Prompts
"Summarise the last 5 PRs in my repo"
"List all open issues in peteroyce/devscope-mcp labelled 'bug'"
"Show me the weekly digest for the myorg/backend repository"
"Who are the top contributors to myorg/frontend this week?"
"Search for usages of
authenticate_useracross the myorg/api repo"
"Give me the full diff and comments for PR #47 in myorg/payments"
"List all repos in the myorg organisation sorted by most recently active"
Running Tests
pytest -v
Tests mock all PyGithub calls so no real GitHub token is required.
Project Structure
devscope-mcp/
├── src/
│ ├── __init__.py
│ ├── github_client.py # PyGithub wrapper — all GitHub logic lives here
│ └── server.py # MCP server — tool definitions, dispatch, formatting
├── tests/
│ └── test_github_client.py
├── .env.example
├── .gitignore
├── pyproject.toml
└── README.md
License
MIT
推荐服务器
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。
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。