AI Code Review Crew

A multi-agent AI system for automated code review. Five specialized agents collaborate sequentially to analyze Python code for bugs, security vulnerabilities, performance issues, and documentation quality.

Note: Live demo not hosted (each review runs 5 LLM agents sequentially — available on request for demonstrations).

🤖 The Agents

1. Code Analyst — Identifies logical errors, edge cases, and exception handling issues
2. Security Expert — Scans for OWASP Top 10 vulnerabilities and security flaws
3. Performance Optimizer — Detects algorithmic bottlenecks and inefficient patterns
4. Documentation Specialist — Reviews docstrings, comments, and code clarity
5. Quality Assurance — Compiles the final report with prioritized recommendations

🎯 Features

• File Upload — Drag-and-drop Python file upload for instant review
• GitHub Integration — Clone and analyze public repositories directly from a URL
• Real-time Agent Status — See which agent is currently analyzing your code
• Detailed Reports — Markdown reports with severity levels, line numbers, and suggested fixes
• Multi-Agent Workflow — Sequential task execution with shared context between agents
• Dark Mode UI — Modern FastAPI interface

🛠️ Tech Stack

• Backend: FastAPI + Python 3.8
• AI Framework: CrewAI (multi-agent orchestration)
• LLM: OpenAI GPT-4
• Frontend: Vanilla JavaScript, HTML, CSS
• Containerization: Docker

🚀 Quick Start

Prerequisites

• Python 3.8+
• OpenAI API key

Installation

1. Clone the repository:

git clone https://github.com/francis-rf/code-review-crew.git
cd code-review-crew

2. Install dependencies:

pip install -r requirements.txt

3. Create .env file:

cp .env.example .env
# Edit .env and add your OPENAI_API_KEY

4. Run the application:

python app.py

5. Open browser:

http://localhost:8000

🐳 Docker Deployment

Build and Run

docker build -t code-review-crew .
docker run -p 8000:8000 --env-file .env code-review-crew

💻 Usage

Upload a File

1. Select Upload File mode
2. Drop your .py file onto the upload area
3. Click Start Code Review
4. Wait for all 5 agents to complete their analysis
5. Download the generated Markdown report

Review a GitHub Repository

1. Select GitHub Repository mode
2. Paste a public GitHub repo URL
3. Select the files to analyze
4. Click Analyze Selected Files

📁 Project Structure

code-review-crew/
├── app.py                  # FastAPI application
├── src/
│   ├── crew.py             # CrewAI orchestration logic
│   ├── logger.py           # Logging configuration
│   └── config/
│       ├── agents.yaml     # Agent definitions (role, goal, backstory)
│       ├── tasks.yaml      # Task definitions
│       └── settings.py     # Application settings
├── static/                 # Frontend
│   ├── index.html
│   ├── app.js
│   └── style.css
├── examples/               # Sample Python files for testing
├── output/                 # Generated review reports
├── logs/                   # Application logs
├── Dockerfile
└── requirements.txt

📡 API Endpoints

Method	Endpoint	Description
GET	/	Serves frontend UI
GET	/health	Health check
POST	/api/review/upload	Review an uploaded Python file
POST	/api/review/github	Review a GitHub repository
GET	/api/files/list	List Python files in a GitHub repository

📸 Screenshots

Code Review Interface showing multi-agent analysis

📄 License

MIT License