Enforces evidence-based reasoning, self-critique, replication validation, and anti-loop detection for all AI agent outputs.
