VulnClaw
AI agent that runs authorized penetration tests from plain language
What it is
VulnClaw is a command-line AI agent that drives penetration testing from recon through vulnerability discovery, exploitation, and report generation based on natural-language goals. It works with OpenAI-compatible and local models, uses an MCP tool chain for fetching, browser automation, and HTTP replay, and includes encoding, payload, and source-recovery helpers plus structured reports with runnable Python PoCs. Built for authorized pentesters, red teams, CTF players, and security instructors, it also offers a TUI and local web UI.