Reverse engineered prompt

Build me a Ricochet Robots solver I can actually use end to end.

I want to be able to run it from the terminal, choose the four board quadrants that make up a normal 16 by 16 board, enter the robot starting positions and target, and get back the shortest solution in a clear move list. Make the core game rules solid, with robots sliding until they hit a wall, obstacle, or another robot.

If the project already has multiple solving approaches, keep the best practical ones working and make sure the result is still optimal. I would also like the random puzzle generator working so it can create lots of valid boards and save solved examples to a csv file.

If possible, keep the Python friendly RL environment usable too, so I can reset, step through moves, and use it for experiments. Please make the whole repo build cleanly, wire up anything that is half finished, and leave simple run instructions so I know how to solve a puzzle and generate data. Look up current docs online if you need to.

Want more depth? Deep Reverse