TRANSITION -> POLICY -> METRIC -> ARTIFACT -> OWNER -> REGRESSION
Intent: domain/state table, policy objects, control polarities
Sequence: save/isolate/off/on/restore/release ordering
Crossings: isolation + level-shifter correctness
Implementation: LP cell placement, rail connectivity, AON integrity
Signoff: simulation + formal + implementation + waiver disciplineWhich transition and domain crossing actually failed first?
Is this policy intent mismatch, sequence ordering bug, or physical issue?
Which artifact proves the hypothesis?
Who owns the minimal reversible fix?
What full LP regression closes release risk?