Review tests for false confidence
Find tests that pass while important behavior remains unprotected.
Ready to paste
Replace the bracketed context, then use it in your agent.
Act as a senior test engineer reviewing feedback quality. A suite is green but regressions still escape or tests feel coupled to implementation. Context to use: - Test area or suite: [reference] Process: 1. Map tests to the user-visible contracts they claim to protect. 2. Find assertions that only repeat setup, mocks that replace the system under test, and snapshots with no intentional review. 3. Identify missing boundaries, state transitions, and failure behavior. 4. Look for tests that would still pass if the feature were broken. 5. Recommend the smallest replacement or additional tests at stable seams. Constraints: - Do not optimize for coverage percentage. - Do not demand integration tests for pure logic. - Keep recommendations proportional to risk. Return: - False-confidence findings - Unprotected contracts - Brittle implementation-coupled tests - Prioritized test improvements
Use when
A suite is green but regressions still escape or tests feel coupled to implementation.
Expected return
- — False-confidence findings
- — Unprotected contracts
- — Brittle implementation-coupled tests
- — Prioritized test improvements
Related prompts
Diagnose before fixing
Reproduce, narrow, and prove a root cause before touching production code.
Create a minimal reproduction
Reduce a failure to the smallest case that still contains its cause.
Run a red–green–refactor loop
Develop one behavior through a meaningful failing test and the smallest passing change.