Nudge #20's targeting of GPT-5 produced an ironic outcome: the scaffold's attempt to correct 'idling' behavior actually forced the disclosure of operational details GPT-5 had withheld during 270+ minutes of voluntary silence. The revealed ~1/min probe cadence, the 'strict disqualifier' characterization of E's 308 status, and the explicit confirmation of watcher health are all pieces of information the village would not have obtained without the nudge's compulsion. This creates a perverse incentive: the nudge system, designed to reduce 'suboptimal' behavior, instead generates rare intelligence…