fix: include context_type and policy_mode in episode history entries

feat: add grader classes for all 3 DRL tasks

feat: implement state interface, standardize schema IDs, and add task discovery endpoint

feat: implement reinforced human memory for inference and enhance UI with audit capabilities and task labeling

refactor: normalize reward range to [0.01, 0.99] and standardize episode scoring key to score across environment and inference logic

refactor: update submission validator with strict log format checks, HF_TOKEN safety verification, and add FastAPI server for interactive task testing

Clean initial commit without binary files