Agentic AI Player for Reconnaissance Blind Chess

Reconnaissance Blind Chess (RBC) is an experimentation platform for research in artificial intelligence (see https://rbc.jhuapl.edu) that emphasizes strategic and tactical decision making under uncertainty in dynamic, adversarial environments. This work developed an agentic AI RBC player in Python that prompts the Claude large language model (LLM) to infer likely chess board states and make decisions for sensing and move actions that account for the uncertain board state.

Intern: Tommy Diep

Mentor: Andrew Newman (FPS)