In my spare time, I vibe-coded (actually, wise-coded) an application that runs multiple AI agents, controlled by a planning AI agent, to perform penetration tests on web applications.
It uses tools like Playwright and Aspire to optimize agentic development.
It's built with .NET and React, but uses alternative components to build and run the spare-time project faster and cheaper.
This session gives you an overview of how I built it and what it uses to run multiple agents for penetration testing.