• Trust-score gaming: Pattern detection; random audits; score rollback; suspension/exit.
• Capture attempts: Power-index concentration alarms; circuit breaker; rotation and caps.
• Sufficiency hoarding: Auto-detection; graduated consequences; possible exit.
• Membrane breaches: Investigation; trust impact; legal action as needed.
• Model failure: Drift detection; revert-to-manual; retrain with human oversight.
• Recovery ladders: Individual → Tetrad → System; always document learning to the pattern library.