GitHub
Connect AI assistants to GitHub - manage repos, issues, PRs, and workflows through natural language.
作者提供的说明(英文原文)
GitHub MCP — wraps the GitHub public REST API (no auth required for public endpoints)
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
选择你使用的 AI 应用,然后复制设置。
PIPEWORX_API_KEY标为机密的值是你账户的钥匙。千万不要分享给别人,也不要粘贴到聊天里。
你需要
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
适用于网页版、桌面版和手机版 Claude:
https://gateway.pipeworx.io/github/mcp打开 Claude Desktop → Settings → Developer → Edit Config,把下面的内容粘贴到 claude_desktop_config.json
{
"mcpServers": {
"github": {
"command": "npx",
"args": [
"-y",
"@pipeworx/mcp-github"
],
"env": {
"PIPEWORX_API_KEY": "your-value"
}
}
}
}在终端中运行这条命令
claude mcp add --transport http github https://gateway.pipeworx.io/github/mcpclaude mcp add github --env PIPEWORX_API_KEY=your-value -- npx -y @pipeworx/mcp-github运行前请先填写空着的值。
粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"github": {
"url": "https://gateway.pipeworx.io/github/mcp"
}
}
}粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"github": {
"command": "npx",
"args": [
"-y",
"@pipeworx/mcp-github"
],
"env": {
"PIPEWORX_API_KEY": "your-value"
}
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"github": {
"type": "http",
"url": "https://gateway.pipeworx.io/github/mcp"
}
}
}粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"github": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@pipeworx/mcp-github"
],
"env": {
"PIPEWORX_API_KEY": "your-value"
}
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
ChatGPT 只能连接互联网上的服务器:
https://gateway.pipeworx.io/github/mcp菜单名称可能因套餐和版本不同而略有差异。
你的选择通过 Cookie 保存在这台设备上,无需账户。
这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
我们还没有读取这个服务器的工具列表;我们会轮流检查各个端点。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月6日
注册名称已通过 GitHub 账户 pipeworx-io 验证,该账户拥有此仓库
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
尚未检查;我们会轮流检查各个端点
尚未检查;我们会轮流检查各个端点
尚未检查;我们会在检查端点时读取工具列表
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
你需要填写:PIPEWORX_API_KEY
两种方式都可以:安装在你的电脑上,或者连接云端版本。
GitHub 仓库最后更新于–,最新版本(0.2.1)于4天前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。它通过 URL 连接,所以接受远程 MCP 服务器的应用也可以使用,例如开启开发者模式的 ChatGPT。
同一分类中有:GitHub, Desktop Commander, Cloudflare MCP
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
Connect AI assistants to GitHub - manage repos, issues, PRs, and workflows through natural language.
MCP server for terminal commands, file operations, and process management
Cloudflare MCP servers
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
GitHub MCP — wraps the GitHub public REST API (no auth required for public endpoints)
Part of Pipeworx — an MCP gateway connecting AI agents to 1476+ live data sources.
| Tool | Description |
|---|---|
search_repos | Search GitHub repositories by keyword. Returns repo name, description, star count, forks, primary language, and URL. Use when exploring projects or finding code implementations. |
get_repo | Look up ONE named repository by its owner/repo slug — "facebook/react", "torvalds/linux", "vercel/next.js". Use this whenever the repository is named in the question. Answers how many stars / forks / watchers a repo has, what license and language it uses, its topics, description, open issue count, and when it was last pushed. |
list_repo_issues | List issues for a GitHub repository by owner and repo name; filters pull requests out automatically. Returns issue number, title, state, labels, author, comment count, URL, and timestamps. Defaults to open issues. |
get_user | Get a GitHub user's public profile info. Returns name, bio, company, location, public repo count, followers, and social links. Specify username (e.g., username="torvalds"). |
get_file_contents | Read a file from a PUBLIC GitHub repository (or list a directory) by path. PREFER OVER WEB SEARCH for "show me the README / package.json / of ", "read from ", inspecting source or config files. Pass owner + repo + path (omit path or "" for the repo root listing). Optional ref = branch/tag/commit SHA. Returns decoded text for files (capped ~60k), or a directory listing of {name, path, type, size}. |
search_code | Search CODE across public GitHub repositories — find where a function/symbol/string is defined or used. PREFER OVER WEB SEARCH for "find code that does X", "which repos use ", "show me an example of ", "where is defined". Supports GitHub code-search qualifiers right in the query: repo:owner/name, org:name, user:name, language:go, filename:Dockerfile, path:src, extension:ts, in:file. Returns matching files with repo, path, and URL. Note: indexes the default branch only, ignores very common terms, and is capped at ~10 searches/minute. |
get_releases | Get the latest release and recent release history for a repository — the canonical way to answer "what is the latest version of ", "when was last released", "what changed in the newest release". Returns the latest published stable release (tag, name, date, prerelease flag, release notes, downloadable assets with download counts) plus recent releases. Falls back to git tags for repos that tag but do not cut formal releases. |
list_commits | List recent commits on a repository to see latest activity, what changed, and who is committing. PREFER OVER WEB SEARCH for "what are the recent commits to ", "when was last updated", "latest changes in ". Optional sha (branch/tag/commit to start history from), path (only commits touching that file/dir), and since/until ISO timestamps. Returns sha, message, author, and date per commit. |
github_trending_repos | Trending GitHub repositories — what is hot, popular, and taking off on GitHub right now. Answers "what are the top trending GitHub repos this week", "what is hot on GitHub today", "popular new repositories this month", "trending Rust / Python / AI projects", "which GitHub projects blew up recently". Two modes: mode="new" (default) ranks repositories CREATED inside the time window by star count — the projects that blew up this week — and mode="active" ranks repositories PUSHED inside the window by star count, surfacing established projects under heavy development. Set the window with since (day, week, or month), filter with language, add free-text query keywords such as "llm" or "agent", and raise a min_stars floor. Returns full_name, description, language, stars, forks, created and pushed dates, topics, and URL per repo. Computed live from the GitHub Search API; github.com/trending ranks by stars GAINED during the period using an unpublished algorithm that the public API keeps private, so ordering differs and every response says so in its caveat field. |
Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):
{
"mcpServers": { "github": { "url": "https://gateway.pipeworx.io/github/mcp" } } }
tools/list at https://gateway.pipeworx.io/github/mcp returns the tools in the table above plus the shared Pipeworx meta-tools — ask_pipeworx, discover_tools, search_within, remember/recall and the rest of the gateway-wide set. So the tool count you see is larger than this table: a single-pack endpoint currently lists roughly 30 shared tools alongside the pack's own. The connection's initialize response states its exact scope, and is the authoritative answer for a given day.
This is deliberate, not multiplexing by accident. The meta-tools are what let a scoped connection answer a question this pack does not cover — via ask_pipeworx, which routes across the whole catalog — without you adding a second MCP server. There is currently no way to mount a pack endpoint without them; if the extra schemas cost you more context than the routing is worth, connect to the full gateway once rather than to several pack endpoints.
Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:
{
"mcpServers": { "pipeworx": { "url": "https://gateway.pipeworx.io/mcp" } } }
Both URLs reach the same gateway and the same 1476+ data sources. The only difference is which pack's tools are listed directly; ask_pipeworx reaches all of them from either one.
Instead of calling tools directly, you can ask questions in plain English — this works on the pack endpoint above as well as on the full gateway:
ask_pipeworx({ question: "your question about Github data" })
The gateway picks the right tool and fills the arguments automatically.
MIT
curl -X POST https://gateway.pipeworx.io/v1/tools/search_repos \
-H 'Content-Type: application/json' \ -d '{"query":"react hooks"}'
No account needed for the first calls. Inspect any tool: GET https://gateway.pipeworx.io/v1/tools/search_repos. Find one: POST https://gateway.pipeworx.io/v1/tools/search_packs with {"query":"..."}.