AI Code Review CLI

🤖 CLI tool for automated code review using local AI models (Ollama).

Features

• 🤖 Local AI Analysis - Uses local AI models via Ollama (privacy-first)
• 🔍 Smart Detection - Finds bugs, performance issues, security vulnerabilities
• 📝 Multiple Modes - Analyze unstaged changes, staged files, or specific commits
• ⚙️ Configurable - Flexible configuration via config files
• 🎨 Beautiful Output - Colored output with icons and severity levels
• 🚀 Multi-language - Supports TypeScript, JavaScript, Python, Go, Rust, Java, C++, and more
• 🔄 CI/CD Ready - Perfect for pre-commit hooks and GitHub Actions

Installation

1. Install Ollama

# macOS/Linux
curl -fsSL https://ollama.ai/install.sh | sh

# Windows - download from https://ollama.ai/

2. Download AI Model

# Recommended model for code review
ollama pull codellama:13b-instruct

# Or lightweight alternative
ollama pull codellama:7b-instruct

3. Install CLI Tool

# Install globally from npm
npm install -g @woobbe/ai-code-review-cli

# Or install locally in project
npm install --save-dev @woobbe/ai-code-review-cli

# Or use npx (no installation required)
npx @woobbe/ai-code-review-cli review

Quick Start

# Start Ollama server (if not running)
ollama serve

# Analyze current changes
ai-code-review review

# Analyze staged changes
ai-code-review review --staged

# Analyze specific commit
ai-code-review review --commit abc123

Usage

Basic Commands

# Analyze unstaged changes
ai-code-review review

# Analyze staged changes
ai-code-review review --staged

# Analyze specific commit
ai-code-review review --commit <hash>

# Quiet mode (errors only)
ai-code-review review --quiet

# Skip health check
ai-code-review review --no-health-check

Configuration Commands

# Show current configuration
ai-code-review config

# Create configuration file
ai-code-review config --init

# Check AI service status
ai-code-review health

Configuration

Create .ai-code-reviewrc.cjs in your project root:

module.exports = {
  modelUrl: 'http://localhost:11434',
  modelName: 'codellama:13b-instruct',
  maxTokens: 4096,
  temperature: 0.05,
  maxFileSize: 1000000, // 1MB
  timeout: 120000, // 2 minutes

  includePatterns: [
    '\\.(ts|tsx|js|jsx|py|go|rs|java|cpp|hpp|cs|rb|php|swift|kt)$'
  ],

  excludePatterns: [
    'node_modules/',
    'dist/',
    'build/',
    '\\.test\\.',
    '\\.spec\\.',
    'package-lock\\.json'
  ],

  reviewPrompt: `You are a senior developer conducting code review. 
Analyze the code and find:

🐛 Bugs and logical errors
⚡ Performance issues  
🔒 Security vulnerabilities
📚 Architectural principle violations
✨ Improvement opportunities
🎨 Code style issues

Focus on critical issues. Be specific and suggest solutions.`
};

Configuration Options

| Option | Description | Default | |--------|-------------|---------| | modelUrl | Ollama server URL | http://localhost:11434 | | modelName | AI model to use | codellama:7b-instruct | | maxTokens | Maximum tokens per request | 2048 | | temperature | AI creativity (0-1) | 0.1 | | maxFileSize | Max file size in bytes | 500000 | | timeout | Request timeout in ms | 120000 | | includePatterns | Files to include (regex) | [] | | excludePatterns | Files to exclude (regex) | Common ignore patterns | | reviewPrompt | Custom AI prompt | Default review prompt |

Examples

Pre-commit Hook

#!/bin/sh
# .git/hooks/pre-commit
ai-code-review review --staged --quiet
exit $?

Make it executable:

chmod +x .git/hooks/pre-commit

Custom Rules

Create specialized configurations for different projects:

// .ai-code-reviewrc.security.cjs - Security focused
module.exports = {
  modelName: 'codellama:13b-instruct',
  reviewPrompt: `You are a security expert reviewing code.
Focus ONLY on:
🔒 Security vulnerabilities
🛡️ Input validation issues
🔐 Authentication/authorization problems
💉 Injection vulnerabilities
🔓 Sensitive data exposure

Rate severity as: critical, high, medium, low.`
};

// .ai-code-reviewrc.performance.cjs - Performance focused
module.exports = {
  modelName: 'codellama:13b-instruct',
  reviewPrompt: `You are a performance optimization expert.
Focus ONLY on:
⚡ Performance bottlenecks
🔄 Inefficient algorithms
💾 Memory leaks
🗄️ Database query optimization
📈 Scalability issues`
};

Output Example

📁 Found 3 files for analysis

📋 Analysis results (3 files):

📁 src/auth.js
  ❌ Using eval() can lead to code injection vulnerabilities (line 12) [security]
    💡 Use JSON.parse() or a safer alternative for parsing data
  ⚠️ Synchronous file operations block the event loop (line 25) [performance]
    💡 Use fs.readFileSync() → fs.promises.readFile() for async operation

📁 src/utils.ts
  ℹ️ Consider using const assertion for better type safety (line 8) [style]
    💡 Change 'as string[]' to 'as const'

📁 src/api.py
  ❌ SQL query vulnerable to injection (line 34) [security]
    💡 Use parameterized queries or ORM methods

📊 Summary: 2 errors, 1 warnings, 1 info

Supported Models

Recommended Models

| Model | Size | Speed | Quality | Use Case | |-------|------|-------|---------|----------| | codellama:7b-instruct | 3.8GB | Fast | Good | Quick reviews, CI/CD | | codellama:13b-instruct | 7.3GB | Medium | Better | Detailed reviews | | codellama:34b-instruct | 19GB | Slow | Best | Comprehensive analysis |

Custom Models

You can use other code-focused models:

• deepseek-coder:6.7b-instruct
• starcoder2:7b-instruct
• codeqwen:7b-instruct

Troubleshooting

Common Issues

Ollama Not Running

# Start Ollama server
ollama serve

# Check if running
curl http://localhost:11434/api/tags

Model Not Found

# List available models
ollama list

# Pull missing model
ollama pull codellama:7b-instruct

# Check model status
ai-code-review health

Slow Analysis

• Use smaller model: codellama:7b-instruct
• Reduce maxTokens in config
• Add more patterns to excludePatterns
• Increase maxFileSize limit

No Changes Detected

# Check git status
git status

# Ensure you're in a git repository
git init

# For unstaged changes
git add -N <new-files>

# For staged analysis
git add <files>
ai-code-review review --staged

Memory Issues

# Monitor Ollama memory usage
docker stats # if using Docker
ps aux | grep ollama

# Use smaller model
ollama pull codellama:7b-instruct

Error Messages

| Error | Solution | |-------|----------| | Current directory is not a Git repository | Run git init or navigate to git repo | | Ollama server unavailable | Start with ollama serve | | Model not found | Download with ollama pull <model> | | Timeout when calling AI model | Increase timeout in config | | File too large for analysis | Increase maxFileSize or exclude file |

Performance Tips

1. Use appropriate model size for your hardware
2. Configure file exclusions to skip unnecessary files
3. Set reasonable timeouts based on model speed
4. Use staged analysis for faster CI/CD pipelines
5. Run health checks before important analysis

Advanced Usage

Custom Prompts for Different Languages

// Language-specific prompts
const prompts = {
  javascript: `Focus on: async/await usage, memory leaks, security vulnerabilities`,
  python: `Focus on: PEP 8 compliance, security issues, performance bottlenecks`,
  rust: `Focus on: unsafe code, borrowing issues, performance optimizations`,
  go: `Focus on: goroutine leaks, error handling, concurrency issues`
};

Integration with IDEs

Use with VS Code tasks:

// .vscode/tasks.json
{
  "version": "2.0.0",
  "tasks": [
    {
      "label": "AI Code Review",
      "type": "shell",
      "command": "ai-code-review",
      "args": ["review", "--quiet"],
      "group": "build",
      "presentation": {
        "echo": true,
        "reveal": "always",
        "focus": false,
        "panel": "shared"
      }
    }
  ]
}

Contributing

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

License

MIT License - see LICENSE file for details.

Acknowledgments

• Ollama for local AI model hosting
• CodeLlama for code understanding
• Contributors and testers

Made with ❤️ for developers who care about code quality