Skip to main content
Glama
repo-scout
by Akash1070

πŸ” Repo Scout MCP Server

"The #1 rule for Vibe Coders and Software Engineers: Don't reinvent the wheel. Let AI scout existing GitHub projects before writing a single line of code!"

A high-performance, free local Model Context Protocol (MCP) server that empowers AI assistants (Google Antigravity, Claude Desktop, Cursor, Claude Code) to search GitHub for existing open-source projects, libraries, and frameworks before you start building from scratch.


πŸ‘€ Author & Credits

  • Author: Akash Kumar Jha

  • License: MIT License (Free & Open Source)


Related MCP server: Implementation Intelligence MCP

πŸš€ Why Use Repo Scout MCP?

When vibe coding or building software, developers often spend hours writing code for features or complete applications that already exist as battle-tested, high-quality open-source projects.

Repo Scout MCP connects your AI coding assistant directly to GitHub's search & intelligence APIs:

  • πŸ’‘ Save Hundreds of Hours: Instantly discover existing solutions before starting a project.

  • πŸ†“ 100% Free & Local: No paid API subscriptions required. Runs locally via Node.js over stdio.

  • ⚑ High Performance: Raises GitHub rate limit from 60 to 5,000 requests/hour with a free token.

  • πŸ›‘οΈ Privacy & Security: Zero hardcoded keys or tracking. Each user supplies their own environment variables.


πŸ› οΈ Included Tools

Tool Name

Description

Example Usage

suggest_stack

Provide a plain-English description of what you want to build. It runs multi-angle GitHub searches (including awesome-lists) and returns a ranked shortlist of repos to reuse.

"I'm building a markdown note app with tags. Check if something exists."

search_repos

Keyword, topic, and language search across GitHub sorted by stars, forks, or recency.

"Find Node.js rate-limiting middleware sorted by stars."

get_repo_details

Pulls a repository's full README and top-level file structure so the AI can evaluate code fit.

"Check details for facebook/react."


πŸ“¦ Quick Start & Installation

Step 1: Clone the Repository

git clone https://github.com/YOUR_USERNAME/repo-scout-mcp.git
cd repo-scout-mcp
npm install

Step 2: Get a Free GitHub Access Token (2 Minutes)

  1. Go to GitHub Settings -> Tokens.

  2. Click Generate new token $\rightarrow$ Generate new token (classic).

  3. Name it repo-scout-mcp.

  4. Leave all scope checkboxes unchecked (public data search requires no permissions).

  5. Click Generate token and copy it (looks like ghp_xxxxxxxxxxxx).


πŸ”Œ Connecting to your AI Assistant

1. Google Antigravity

  1. Open Antigravity $\rightarrow$ click the ... menu at top-right of agent panel.

  2. Go to MCP Servers $\rightarrow$ Manage MCP Servers $\rightarrow$ View raw config (mcp_config.json).

  3. Add the following entry inside "mcpServers":

{
  "mcpServers": {
    "repo-scout": {
      "command": "node",
      "args": ["G:\\CODE\\repo-scout-mcp\\index.js"],
      "env": {
        "GITHUB_TOKEN": "ghp_your_personal_access_token_here"
      }
    }
  }
}

(Note: Replace G:\\CODE\\repo-scout-mcp\\index.js with your actual full file path, using double backslashes on Windows).

  1. Save mcp_config.json and refresh MCP servers (or type /mcp in chat).


2. Claude Desktop

Add to your claude_desktop_config.json:

  • Windows: %APPDATA%\Claude\claude_desktop_config.json

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "repo-scout": {
      "command": "node",
      "args": ["/absolute/path/to/repo-scout-mcp/index.js"],
      "env": {
        "GITHUB_TOKEN": "ghp_your_personal_access_token_here"
      }
    }
  }
}

3. Claude Code CLI

Run:

claude mcp add repo-scout node /absolute/path/to/repo-scout-mcp/index.js -e GITHUB_TOKEN=ghp_your_token_here

πŸ’¬ Example Prompts to Try

Once connected, simply prompt your AI assistant:

  • Before starting a project:

    "I'm planning to build a self-hosted dashboard for monitoring Docker containers. Use repo-scout to check if a great open-source project already exists."

  • Searching for libraries:

    "Use repo-scout to search for high-star TypeScript financial technical indicator packages."

  • Evaluating a repository:

    "Use repo-scout get_repo_details on shadcn/ui to summarize its structure."


πŸ“„ License

Created by Akash Kumar Jha. Released under the MIT License.

Available Tools

3 tools
get_repo_detailsA

Fetch details for one specific repo (owner/name), including README excerpt and top-level file list, so you can judge whether it fits your use case.

Inputs: repo (required).

search_reposB

Search public GitHub repositories by keyword/topic. Returns stars, last-updated, license, etc.

Inputs: query (required), sort, limit, language, +1 more.

suggest_stackA

Given a plain-language description of what you're about to build, runs several targeted GitHub searches (including curated/awesome-list style queries) and returns a ranked shortlist of existing projects/libraries you could reuse instead of building from scratch.

Inputs: description (required), language, limit_per_query.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A4/5.0

    Scored across 3 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose: retrieving details for a specific known repo, searching by keywords, and generating a shortlist from a natural-language description. No overlap in functionality or target resource.

    Naming Consistency5/5

    All tool names follow a consistent verb_noun pattern in snake_case: get_repo_details, search_repos, suggest_stack. This is a predictable and uniform convention.

    Tool Count5/5

    With only 3 tools, the server is tightly scoped to the 'repo scout' purpose: discovery (search), semantic suggestion (suggest), and evaluation (get details). Each tool earns its place, and the count is within the ideal 3-15 range.

    Completeness5/5

    The tool surface covers the full workflow for scouting repositories: finding candidates (search_repos, suggest_stack) and inspecting them (get_repo_details). No obvious gaps or dead ends for the stated purpose.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      A
      quality
      D
      maintenance
      Enables AI agents to prospect across GitHub, PyPI, and npm, searching repositories, packages, and code patterns with read-only tools.
      7
      80 PyPI
      5
      MIT
    • A
      license
      Not graded
      quality
      B
      maintenance
      Enables AI coding agents to discover, evaluate, and select the best existing open-source implementations by assessing license compatibility, extracting only the needed symbols and dependencies, and producing integration or adaptation plans.
      Apache 2.0
    • A
      license
      A
      quality
      C
      maintenance
      Enables AI agents to analyze public GitHub repositories directly through tools for repo health scoring, recent commit summaries, good first issue discovery, and semantic Q&A over docs and issues.
      4
      31 npm
      1
      MIT
    • A
      license
      A
      quality
      C
      maintenance
      Enables AI assistants to search, inspect, compare, and evaluate GitHub repositories using read-only tools, including topic discovery, repository health scoring, security advisories, dependency audits, and supply-chain risk checks. It returns concise, readable results and can work without a token, with optional authentication for higher rate limits and code search.
      16
      62 npm
      MIT