BOUNDARY -> MODE -> MECHANISM -> METRIC -> ARTIFACT -> OWNER -> REGRESSION
Boundaries: cells, wrappers, reset and CDC assumptions
Clocking: lock quality, jitter budget, uncertainty impact
High-speed I/O: training state, BER, lane margin, SI context
Converters: sample timing, calibration hooks, validity flags
Physical: keepouts, guard/substrate, package and IR/noise
Signoff: dashboard blockers, waivers, owner approvalsWhich boundary fails first under which mode?
Which mechanism explains that failure?
Which metric and artifact prove it?
Who owns smallest reversible fix?
Which regressions are mandatory before release?