PRobe

Deep scan. Smart review. Ship with confidence.

PRobe is an AI-powered GitHub Pull Request analysis dashboard that helps maintainers and tech leads make informed code review decisions. Paste a GitHub repo URL, select any open PR, and get instant insights powered by IBM Watson Machine Learning with Granite 3.0 8B Instruct.

🎯 What PRobe Does

PRobe analyzes pull requests and surfaces:

• Plain English Summaries: Understand what the PR does in seconds, written for non-technical stakeholders
• Refactor Score: Visual circular progress ring showing code quality (0-100 scale)
• Test Coverage Analysis: Identifies functions missing test coverage with severity ratings
• Security Scanning: Detects vulnerabilities, hardcoded secrets, and insecure patterns
• PR Size Warnings: Flags oversized PRs and mixed concerns with split suggestions
• Auto-Label Suggestions: AI-generated labels with confidence scores

All analysis happens in one single AI call to conserve credits and maximize efficiency.

🚀 Features

For Maintainers

• Zero Configuration: Works with public GitHub repos, no authentication required
• Instant Analysis: Results in ~10-15 seconds per PR
• Smart Caching: Analyzed PRs are cached for 30 minutes
• Brutalist UI: Clean, distraction-free interface with true black (#0a0a0a) background

For Tech Leads

• Risk Assessment: Refactor scores highlight technical debt
• Security First: CVE detection, secret scanning, and pattern analysis
• Review Efficiency: Plain English summaries save review time
• Data-Driven Decisions: Confidence scores on all AI suggestions

🛠️ Tech Stack

• Frontend: Next.js 14+ (App Router), React, TypeScript
• Styling: Tailwind CSS + Custom Brutalist Design System
• AI Engine: IBM Watson Machine Learning (Granite 3.0 8B Instruct)
• GitHub Integration: Octokit (GitHub REST API v3)
• Deployment: Vercel-ready with zero-config deployment

📦 Installation

Prerequisites

• Node.js 18+ and npm
• IBM Watson Machine Learning credentials
• (Optional) GitHub Personal Access Token for private repos

Setup

1. Clone the repository

   git clone https://github.com/yourusername/probe.git
   cd probe

2. Install dependencies

   npm install

3. Configure environment variables

   Create a .env.local file in the probe directory:

   # IBM Watson Machine Learning
   WATSON_ML_URL=https://us-south.ml.cloud.ibm.com
   WATSON_ML_API_KEY=your_watson_api_key_here
   WATSON_ML_MODEL_ID=ibm/granite-3-8b-instruct
   WATSON_ML_PROJECT_ID=your_project_id_here
   
   # GitHub (optional, for private repos)
   GITHUB_TOKEN=your_github_token_here

4. Run the development server

   npm run dev

5. Open your browser

   Navigate to http://localhost:3000

Design System

PRobe uses a custom brutalist design system:

• Colors: True black (#0a0a0a), Electric Violet (#8b5cf6) accent
• Typography:
  • Bebas Neue (display numbers)
  • Space Grotesk (headings)
  • DM Mono (code/metadata)
  • Inter (body text)
• Layout: 2-column grid (220px sidebar + flexible main)
• Components: Borderless sections with subtle dividers

🔧 Configuration

Watson ML Setup

1. Create a Watson Studio project in IBM Cloud
2. Note your Project ID from the "Manage" tab
3. Generate an API key from IBM Cloud IAM
4. Add credentials to .env.local

GitHub Token (Optional)

For analyzing private repositories:

1. Go to GitHub Settings → Developer settings → Personal access tokens
2. Generate a token with repo scope
3. Add to .env.local as GITHUB_TOKEN

📖 Usage

Analyzing a PR

1. Enter a GitHub repo URL

   https://github.com/owner/repo
   

   or just owner/repo

2. Click "LOAD PRS →"

   All open PRs will appear in the left panel

3. Select a PR

   Click any PR card to trigger analysis

4. Review the insights

   • Summary (top right)
   • Refactor Score (left sidebar)
   • Test Coverage, Security, PR Size, Labels (right column)

Understanding the Refactor Score

• 0-39 (Green): Code is clean, no refactor needed
• 40-74 (Amber): Refactor optional but beneficial
• 75-100 (Red): Strongly consider refactoring

The score considers:

• Cyclomatic complexity
• Function length (>50 lines penalized)
• Code duplication
• Single Responsibility Principle violations
• Nesting depth (>4 levels penalized)

Project Structure

probe/
├── app/
│   ├── page.tsx              # Landing page
│   ├── dashboard/
│   │   └── page.tsx          # Main dashboard
│   ├── api/
│   │   ├── prs/route.ts      # Fetch PRs endpoint
│   │   └── analyze/route.ts  # AI analysis endpoint
│   ├── layout.tsx            # Root layout
│   └── globals.css           # Global styles
├── components/
│   ├── TopBar.tsx            # Logo + input + button
│   ├── PRList.tsx            # Left panel PR list
│   ├── PRCard.tsx            # Individual PR cards
│   ├── AnalysisPanel.tsx     # Right panel layout
│   ├── BobThinking.tsx       # Loading state
│   ├── analysis/
│   │   ├── SummarySection.tsx
│   │   ├── RefactorScore.tsx
│   │   ├── TestCoverage.tsx
│   │   ├── SecuritySection.tsx
│   │   ├── PRSizeWarning.tsx
│   │   └── LabelSuggestions.tsx
│   └── ui/
│       ├── Card.tsx
│       ├── Badge.tsx
│       ├── Button.tsx
│       └── Input.tsx
├── lib/
│   ├── bob.ts                # Watson ML integration
│   ├── github.ts             # GitHub API wrapper
│   ├── cache.ts              # In-memory cache
│   └── utils.ts              # Helper functions
├── types/
│   └── index.ts              # TypeScript interfaces
└── public/                   # Static assets

Deployment

Vercel (Recommended)

1. Push your code to GitHub
2. Import the project in Vercel
3. Add environment variables in Vercel dashboard
4. Deploy

Other Platforms

PRobe is a standard Next.js app and can be deployed to:

• Railway
• Render
• AWS Amplify
• Netlify
• Self-hosted with Docker

Contributing

Contributions are welcome! Please follow these steps:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• IBM Watson Machine Learning for providing the Granite 3.0 8B Instruct model
• GitHub for the comprehensive REST API
• Vercel for Next.js and deployment platform
• Tailwind CSS for the utility-first CSS framework

Support

For issues, questions, or feature requests:

• Open an issue on GitHub Issues

IBM Bob Hackathon 2026

PRobe was built for the IBM Bob Hackathon 2026 to explore how AI development partners like IBM Bob can transform modern code review workflows.

By combining contextual code understanding with intelligent pull request analysis, PRobe helps engineering teams review changes faster, identify potential risks earlier, and ship with greater confidence.

The project reflects the core vision of IBM Bob: AI that works with real repositories, understands development context, and assists across practical software engineering workflows.

Built with ❤️ using IBM Granite 3.0 8B Instruct