causal-bug-proof
Use when a bug needs stronger evidence than an ordinary regression test: a regression after a release, upgrade, refactor or merge; a cause that is unclear or disputed between your code, a dependency, the environment and a flaky test; a test suite that stays green while the bug is live; or a fix that works without anyone knowing why. Triggers include "find the root cause", "prove this fix", "which commit broke this", "why didn't the tests catch it" and "is this our bug or upstream's". Pins the wanted behaviour in a failing test whose assertions never change, traces the code path and the git history, applies the smallest fix, and shows that reverting only the fix brings the failure back. Prevents fixes that pass by coincidence, tests weakened to fit old code, and root-cause claims stronger than the evidence.
