hackaton-mcp-server

Pro-dutt/hackaton-mcp-server

3.3

If you are the rightful owner of hackaton-mcp-server and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

A lightweight Model Context Protocol (MCP) server built for hackathons, enabling rapid prototyping of AI-powered applications.

🔍 Repository Discovery Platform

MCP Hackathon 2025 - Theme 1: Civic Engagement

Repo Scout

A smart repository discovery platform that helps developers find the perfect GitHub repositories for their projects and learning journey.

🎯 Problem Statement

In today's vast open-source ecosystem, developers struggle to:

  • Find relevant repositories for new projects
  • Discover beginner-friendly repositories to contribute to
  • Get meaningful brainstorming ideas from existing codebases
  • Identify repositories that match their skill level and interests

💡 Solution Overview

Our platform leverages Model Control Protocol (MCP) servers and AI to provide intelligent repository recommendations through a user-friendly web interface.

Key Features

  • Smart Repository Search: AI-powered search using GitHub MCP tools
  • Beginner-Friendly Filtering: Identifies good first issues and beginner resources
  • Repository Analysis: Deep analysis of repository quality, activity, and community
  • Comparative Analysis: Compare multiple repositories to make informed decisions
  • Personalized Recommendations: Tailored suggestions based on user preferences

🏗️ Architecture

Two Implementation Approaches

Approach 1: LM Studio Integration
  • Primary Stack: Next.js frontend + Flask backend
  • AI Model: LM Studio with customizable parameters
  • MCP Integration: Docker-hosted GitHub MCP servers
  • Communication: CLI streams with stdin/stdout
  • Orchestration: LangGraph for workflow management
Approach 2: Claude Desktop (Optimized)
  • Platform: Claude Desktop application
  • Performance: Enhanced speed and efficiency
  • MCP Tools: Direct integration with GitHub and custom repo_analyzer MCPs

🛠️ Tech Stack

Frontend

  • Framework: Next.js
  • Authentication: Descope (OAuth with GitHub & Google)
  • Styling: Modern responsive design

Backend

  • API Server: Flask
  • AI Integration: LM Studio API / Claude Desktop
  • Workflow: LangGraph for repository filtering and analysis
  • Containerization: Docker for MCP servers

MCP Tools

  1. GitHub MCP (Standard)

    • search_repository
    • get_issues
    • get_directory_contents

  2. Custom repo_analyzer MCP (FastMCP)

    • analyze_repository
    • get_beginner_resources
    • suggest_good_first_issues
    • compare_repositories

🚀 Getting Started

Prerequisites

  • Node.js 18+
  • Python 3.8+
  • Docker
  • LM Studio (for Approach 1)

Installation

  1. Clone the repository

    git clone <repository-url>
cd repository-discovery-platform

  2. Frontend Setup

    cd frontend
npm install
npm run dev

  3. Backend Setup

    cd backend
pip install flask httpx fastmcp flask-cors
python app.py

  4. MCP Server Setup

    docker-compose up -d

Configuration

  1. Environment Variables

    GITHUB_PERSONAL_ACCESS_TOKEN = <your_github_token>
DESCOPE_PROJECT_ID = your_descope_project_id
DESCOPE_MANAGEMENT_KEY = descope_management_key
LM_STUDIO_API_URL = http://localhost:1234

  2. LM Studio Configuration

    • Start LM Studio server on port 1234
    • Load your preferred model
    • Configure MCP integration

📱 Usage

Web Interface

  1. Sign In: http://localhost:3000/sign-in

    • GitHub or Google OAuth via Descope

  2. Generate Recommendations: http://localhost:3000/generate

    • Enter your project description or interests
    • Customize AI parameters (temperature, max tokens, top-k, etc.)
    • Get intelligent repository recommendations

API Endpoints

LM Studio Endpoints
  • GET /health - Server health check
  • GET /models - Available models
  • POST /chat - Process prompts
  • POST /chat/completions - Get AI completions

🔧 Features Deep Dive

Intelligent Filtering

  • Star Count Analysis: Repository popularity metrics
  • Commit History: Activity and maintenance status
  • Issue Management: Active community engagement
  • Contributor Analysis: Community size and diversity

Repository Quality Assessment

  • README Quality: Documentation completeness
  • License Compliance: Open-source license verification
  • Beginner Friendliness: Good first issues and contribution guides
  • Project Structure: Code organization and best practices

Personalization

  • Skill Level Matching: Beginner to advanced recommendations
  • Technology Stack: Language and framework preferences
  • Contribution History: Based on user's GitHub activity

🎮 Demo

Architecture

🏆 Hackathon Results

Achievements

  • Successfully integrated multiple MCP servers
  • Created custom MCP tools for enhanced repository analysis
  • Implemented two different architectural approaches
  • Achieved significant performance improvements in Approach 2

Lessons Learned

  • MCP integration challenges with stdin/stdout communication
  • Performance optimization through Claude Desktop integration
  • Importance of user experience in AI-powered applications

👥 Team

🙏 Acknowledgments

  • MCP Hackathon organizers
  • GitHub API for repository data
  • LM Studio team for local AI model hosting
  • Descope for authentication solutions
  • FastMCP for simplified MCP development

Built with ❤️ for the MCP Hackathon 2025