AmulyaJain2004/HackOpenEnv ? reverse-engineered prompt

Build me a working reinforcement learning environment called Bureaucracy Escape Room, with the main environment code, a simple inference script, Docker support, and validation scripts so I can run it locally and submit it for the hackathon.

I want the repo to be easy to set up from the root with the requirements file, then go into the environment folder, build the Docker image, and run a baseline inference without a lot of manual tweaking. Please make sure the environment is actually usable and that the validation scripts can check a submission before I send it in.

If anything depends on current API or Docker behavior, look up the latest docs online if needed. Keep the setup straightforward, make the scripts work cleanly on a normal machine, and include any missing pieces needed so the environment can be built, run, and validated end to end.