Regression Risk Hunter
Built for the PR that looks trivial in the diff but quietly changes an assumption something downstream relies on. - Flags changed assumptions, edge cases, and null/empty handling - Checks off-by-one logic, state transitions, and API contract drift - For each issue: what changed, what it could break, and the smallest test that would have caught it
Prompt
Review this PR for regressions. Focus on: - changed assumptions - edge cases - null/empty behavior - off-by-one logic - state transitions - API contract drift For each issue, explain: - what changed - what might break - the smallest test that would catch it
