ai-regression-testinglisted

# CIEL ADAPTATION: AI Regression Testing This skill provides CIEL's specialized testing strategy for mitigating systematic blind spots in AI-assisted development. It focuses on contract-first verification and sandbox/production parity. ## Integration Context Adapted from `~/.agents/skills/ai-regression-testing/`. This skill complements the core **TDD** mandate by targeting common AI failure patterns, such as "hallucinated success" and path inconsistency. ## Core Patterns ### 1. Sandbox/Production Parity **Problem**: AI fixes the production path but forgets the sandbox/mock path (or vice versa). **Solution**: Assert identical response shapes in both modes. ### 2. Contract Verification **Problem**: AI adds a field to the response object but forgets to include it in the underlying data query (e.g., SELECT clause). **Solution**: Assert all required fields exist in the final API response. ### 3. Error/State Integrity **Problem**: old data persists in the UI after an error, or optimistic updates lack proper rollback. **Solution**: Explicit tests for state cleanup and restoration on failure. ## Orchestration Logic ### 1. Post-Fix Verification After every bug fix (via **systematic-debugging**), CIEL must generate a specific regression test named after the bug (e.g., `BUG-R1 regression`). ### 2. Bug-Check Workflow Run `npm run test` and `npm run build` **before** any AI-led code review to catch mechanical failures early. ### 3. Strategy: Test Where Bugs Cluster Focus