Transform vulnerability write-ups into fully functional, runnable security labs

LabGenie is a terminal-first, interactive CLI tool that takes a vulnerability write-up URL and produces a complete, runnable lab environment ready for security research and education.

• Python 3.10 or higher
• Option A: Google Gemini API key (Get one here) - Recommended
• Option B: Google Cloud Project with Vertex AI enabled

1. Clone or navigate to this directory

cd /path/to/LabGenie

2. Run the setup script

bash setup.sh

This will:

• Create a Python virtual environment
• Install all dependencies
• Check for required API keys
• Make the CLI executable

3. Configure your AI provider

Option A: Gemini API (Recommended)

export GOOGLE_API_KEY='your-gemini-api-key'

Option B: Vertex AI (Enterprise)

export GOOGLE_CLOUD_PROJECT='your-gcp-project-id'
gcloud auth application-default login

Or create a .env file:

cp .env.example .env
# Edit .env and add your API key

# Activate virtual environment (if not already active)
source venv/bin/activate

# Interactive mode (default)
python labgenie.py

# Direct URL mode
python labgenie.py --url https://example.com/vuln-writeup

# Debug mode for detailed output
python labgenie.py --debug

Simply run the CLI and provide a vulnerability write-up URL:

$ python labgenie.py

🔗 Write-up URL: https://example.com/blog/sqli-vulnerability

The CLI will then:

1. Convert the write-up to structured markdown
2. Parse vulnerability information and reproduction steps
3. Plan the lab architecture and components
4. Build complete, runnable lab artifacts

All generated files are saved to ./output/{labname}/:

output/
└── sqli_vulnerability_lab/      # Lab name extracted from vulnerability
    ├── lab_manifest.json         # Complete lab metadata
    ├── README.md                 # Lab setup instructions
    ├── docker-compose.yml        # Container orchestration
    └── src/                      # Application source code

LabGenie implements a four-stage AI agent workflow. For detailed information about the system architecture, agent designs, and technical implementation, see:

docs/Architecture.md

• Architecture - System design and agent workflow
• Troubleshooting - Common issues and solutions

Contributions are welcome!

LabGenie is inspired by the paper From CVE Entries to Verifiable Exploits. Shoutout to the researchers at UC Santa Barbara who created it—it's a wonderful multi-agent framework that generates POC code from CVE entries. Their approach inspired LabGenie, which tackles a related challenge from a different angle. Huge thanks to the contributors of the whitepaper for lighting the way.

Made with 🧞 magic by LabGenie