github-mcp-lite
A lightweight MCP server for GitHub with nine essential tools for reading files, managing pull requests, and creating issues, using personal access token authentication.
README
github-mcp-lite
A small Model Context Protocol server for GitHub — nine tools, PAT auth, stdio transport.
The official github/github-mcp-server covers a lot of ground (60+ tools). Handy for org admin and security workflows, but heavy if you mostly read files, scan PRs, and file issues. GitHub's been nudging people toward smaller tool sets for that reason (changelog, notes on token use).
This repo is the opposite end: fixed scope, short tool list, easy to skim before you plug it into an MCP host.
Install
git clone https://github.com/benjimkh/github-mcp-lite.git
cd github-mcp-lite
python3 -m venv .venv
source .venv/bin/activate
pip install -e .
Run it:
export GITHUB_TOKEN=ghp_...
github-mcp-lite
MCP client config (stdio):
{
"mcpServers": {
"github-mcp-lite": {
"command": "/absolute/path/to/github-mcp-lite/.venv/bin/github-mcp-lite",
"env": {
"GITHUB_TOKEN": "ghp_your_token_here"
}
}
}
}
Tools
| Tool | What it does |
|---|---|
get_repo |
Metadata, default branch, visibility, star count |
search_code |
Code search (repo:owner/name queries work well) |
get_file_contents |
Read a file or list a directory at a ref |
list_pull_requests |
Open / closed / all PRs |
get_pull_request |
Single PR with merge state and diff stats |
create_issue |
Open a new issue |
comment_on_issue |
Comment on an issue or PR thread |
list_issues |
Issues only (PRs filtered out) |
search_repositories |
Repo search |
No gists, notifications, or org settings — that's intentional.
Benchmark
Ballpark tool/schema footprint vs the official server:
python3 benchmark/measure_schema_size.py
Background: docs/context-budget-guide.md.
This repo vs official github-mcp-server
| github-mcp-lite | official server |
|---|---|
| Few repos, everyday PR/issue work | Org admin, security, Dependabot |
| Short tool list you can read in a minute | Actions, Projects, Discussions, etc. |
| PAT + stdio | Copilot OAuth integration |
License
MIT — see LICENSE. PRs welcome; see CONTRIBUTING.md.
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。