igorganapolsky

Pre-action gates that block AI agents from repeating known mistakes.