Hostile-by-default critic that grades AI agent output against your acceptance policy.
Probing this server's capabilities…