Regression bisect
Pin a regression to the exact change that introduced it.
The prompt
This worked before my recent changes. Review the diff in @<files> and identify which specific change most likely caused <symptom>. Explain the mechanism.Why it works
Scoping the search to your recent diff turns “somewhere in the codebase” into a small, reviewable set. Requiring a mechanism (not just a guess) keeps the model honest.
Fill in
@<files>— the files you changed (or pastegit diff).<symptom>— what broke.
Follow-ups
- “Revert just that hunk and tell me if the logic still holds.”