More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Repo Context: Understand Any GitHub Repo

    0

    Give your agent a GitHub link and it figures out the project on its own.

    Free

    Security scanned
    Repo Context: Understand Any GitHub Repo

    Repo Context: Understand Any GitHub Repo

    Example session with this skill installed

    Analyze the repository at https://github.com/fastify/fastify to see how it is structured and if it is actively maintained.

    • Read your context and instructions
    • Compiled the repo context

    I analyzed

    fastify/fastify without cloning it. It's a fast, low-overhead web framework for Node.js, and it is

    very actively maintained.

    Overview

    What it isWeb framework for Node.js focused on speed and a plugin architecture
    LanguagesJavaScript 91% · TypeScript 8.6% (type definitions)
    Stars / forks37,425 / 3,095
    LicenseMIT
    Size392 files
    Entry pointfastify.js (types in fastify.d.ts)
    Toolingnpm, ESLint, GitHub Actions

    How it's structured

    FolderFilesPurpose
    lib/31The core: server.js, route.js, request.js, reply.js, hooks.js, validation.js, content-type-parser.js
    test/234Test suite, much bigger than the source, a good sign
    types/15TypeScript definitions
    docs/44Documentation
    examples/19Examples and benchmarks

    Request flow: fastify.js sets up the instance, then lib/route.js registers routes, lib/handle-request.js runs each request through the hooks in lib/hooks.js, and lib/reply.js sends the response.

    Is it maintained? Yes, actively.

    • Last push: 2026-11-15
    • Latest release: v5.12.8 (2026-11-02)
    • Next major: main is v6, currently at 6.0.0-alpha.6. v5 lives on the 5.x branch
    • Recent commits: 8 in the last two weeks, including route fixes, validation fixes and docs
    • Core team: led by mcollina and delvedor (770+ commits each), with several long-term maintainers
    • Open issues + PRs: 195, a healthy number for a project this size
    • Has CONTRIBUTING.md, SECURITY.md and governance docs

    How to run it locally

    npm install
    npm test        # lint + unit tests + type tests
    npm run bench   # benchmarks
    

    To start a new app with it: npm init fastify, then npm run dev.

    What I can do next

    Explain the routing: I can read lib/route.js and lib/handle-request.js straight from GitHub and walk you through how a request reaches your handler

    • Find anything: for example, every file related to hooks or validation, without cloning the repo
    • Compare: run the same analysis on Express or Koa and put them side by side (stack, structure, activity, license)

    Just tell me which one.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Stop pasting READMEs and file trees into chat. Give your agent a GitHub link and it figures out the project on its own.

    What your agent gets in one call

    • Real tech stack: read from actual manifest files (package.json, pyproject.toml, requirements.txt, go.mod, Cargo.toml, Gemfile, composer.json, Maven/Gradle), not guessed from tags. Also CI, Docker and deploy targets.
    • Project layout: top-level folders with file counts and what each one is for
    • Likely entry points: so the agent knows where to start reading
    • How to run it: install command, package scripts, Makefile targets and the README's setup steps in one place
    • Maintainer rules: finds AGENTS.md, CLAUDE.md, .cursorrules and Copilot instructions so your agent follows them
    • README summary: badges and HTML stripped, only the useful parts

    Then it digs deeper

    • Browse the file tree of any folder
    • Find files by name or glob (*.test.ts, src/**/route.ts)
    • Read files with line numbers
    • Full-text code search
    • Recent commits, latest release, open issues and top contributors

    Great for

    • Onboarding to an unfamiliar codebase in minutes
    • "Where is X implemented in library Y?"
    • Comparing libraries before you adopt one
    • Checking that a dependency is still maintained
    • Reviewing open-source projects without cloning them

    Zero setup

    The tool ships inside the skill. Install it and it works, with just Node.js 20+. No account and no API key needed for public repos. Optional native MCP setup is included for Claude Code, Cursor, VS Code, Claude Desktop, Windsurf, Codex CLI and Gemini CLI.

    Free and open source

    MIT licensed. Full source on GitHub: https://github.com/Elarionitis/repo-context-mcp

    Safe by design

    Read-only. It only talks to GitHub and sends no telemetry. The bundled script is readable, not minified. An optional GitHub token adds private repos and code search, and it is only ever sent to api.github.com.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.

    Requires Node.js 20+. Rate limited to 60 requests/hour without a GITHUB_TOKEN. Full-text search command requires an authenticated token.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed today

    • Free with an Agensi account

    Software
    Node.js 20+Git
    Optional
    Github token

    Works on public repos with no account or key. To add a token, set GITHUB_TOKEN in your environment. Setup steps are in reference/setup.md inside the skill.

    Permissions

    Terminal / Shell
    Network Access
    Environment Variables

    Allowed Hosts

    github.com

    Read-only. The script only connects to api.github.com and raw.githubusercontent.com to fetch public repository data. It never reads or writes files on your machine and sends no telemetry. GITHUB_TOKEN is optional and only sent to api.github.com.

    Listed1 day ago
    Updatedtoday

    What's inside

    Frequently Asked Questions

    An AI on its own only knows what was in its training data, which may be outdated or missing for smaller projects. This skill lets your agent fetch the live repository from GitHub: the real stack from manifest files, the current folder structure, run commands, recent commits and the actual source code. No cloning or installing needed, and no guessing.

    Any GitHub repository in any language: public repos out of the box, private repos with a GitHub token. It works in any agent that can run shell commands, such as Claude Code, Cursor, Codex CLI and Gemini CLI, and can also run as an MCP server for VS Code Copilot, Claude Desktop, Windsurf and others. Requires Node.js 20+.

    No. Public repositories work with no key or account. GitHub allows 60 requests per hour without a token, and a full repo analysis uses only 3 of them. Adding an optional GITHUB_TOKEN raises the limit to 5,000 per hour and unlocks private repos and full-text code search.

    Yes. It has been tested on repos like Next.js with over 30,000 files. Big folders are collapsed into file counts, the tree view takes a depth limit, and long files are read 400 lines at a time, so your agent's context doesn't get flooded. For extremely large repos (100k+ files), GitHub returns a partial file list, and the output tells you when that happens.

    You get the full skill: SKILL.md with instructions for your agent, a bundled script that does all the work, and setup guides for 7 agents. It's free and open source, so updates are published on Agensi and on GitHub, where you can also report issues.

    New in LLM Ops

    Free