Agents that reason, act, and learn inside a controlled, virtual environment to experiment and fail safely while refining strategy before anything touches a real-world environment. - View it on GitHub