> The agents missed it while going through the literature

Is this verifiable? I wonder given the recent publications regarding intentional deception that is ultimately opaque to end users.