

Most AI security CTFs are closed source, so you can’t even be sure it works. That’s why we open-sourced Playground. Every challenge uses a live agent with real tools, and we publish its prompt, configuration, defender code, and successful attacks. Show us your hacking skills!










