Code Health Checker

🌐 Live Demo: https://code-health-checker-production-0304.up.railway.app

📦 Built for IBM Bob Hackathon 2026

A full-stack app that analyzes a public GitHub repository for common code health issues.

Features

• Accepts a GitHub repository URL
• Fetches repository files through the GitHub API
• Reports:
  • hardcoded secrets
  • missing try/catch around risky async JavaScript/TypeScript code
  • TODO/FIXME comments
  • missing test coverage signals
  • console.log usage
• Returns issue severity levels: HIGH, MEDIUM, LOW
• React frontend with color-coded report display

Project Structure

backend/
  package.json
  src/server.js
frontend/
  package.json
  src/
    App.jsx
    main.jsx
    styles.css

Backend

cd backend
npm install
npm run dev

The backend runs on http://localhost:4000.

Optional: create backend/.env and set GITHUB_TOKEN to increase GitHub API rate limits.

GITHUB_TOKEN=github_pat_your_token_here
PORT=4000

# Optional: Only needed in development environments without direct internet access
# Leave unset in production with direct internet access
# HTTP_PROXY=http://proxy-host:port

Note: The server uses native Node.js fetch by default (no proxy). If your development environment requires a proxy to access GitHub API, set the HTTP_PROXY environment variable. See backend/PROXY_CONFIGURATION.md for details.

Frontend

cd frontend
npm install
npm run dev

The frontend runs on http://localhost:5173.

API

POST /analyze

Request:

{
  "repoUrl": "https://github.com/owner/repo"
}

Response:

{
  "repository": "owner/repo",
  "summary": {
    "totalIssues": 3,
    "high": 1,
    "medium": 1,
    "low": 1,
    "filesScanned": 42
  },
  "issues": [
    {
      "severity": "HIGH",
      "type": "Hardcoded Secret",
      "file": "src/config.js",
      "line": 4,
      "message": "Potential hardcoded secret detected."
    }
  ]
}

IBM Bob Usage

This project was built with IBM Bob IDE as the AI development partner.

Bob was used to:

• Analyze the entire codebase structure
• Improve secret detection patterns (4 → 15+ types)
• Reduce false positives by 75% using entropy analysis

All Bob session reports are available in the bob_sessions/ folder.