What problem does it solve? Failing tests often get patched with quick fixes like sleeps, mocks, or re-runs that hide the real defect. This Skill enforces a structured debugging workflow that classifies the true root cause of a test failure before any application code is touched. ## Core Features & Use Cases - Root-Cause Classification: Investigates failures in a fixed order: fixtures and conftest configuration, environment mismatches (SQLite vs Postgres, env vars), dependency drift, test isolation issues, and only then application logic. - Reproduce-Fix-Verify Workflow: Captures full pytest output, applies a minimal targeted fix, re-runs the failing test in isolation, then runs the full suite to check for side effects. - uv Environment Stability: Documents how to pin a uv environment with uv sync --all-extras and uv run --no-sync so iterative debug runs do not silently drop extras and cause misleading ModuleNotFoundError failures. - Use Case: A CI pipeline reports a failing auth test that passes locally. Invoke the skill with the test path to reproduce the failure, classify it as an environment mismatch, and fix the missing env var instead of mocking the failure away. ## Quick Start Ask the AI to debug the failing test at tests/unit/test_auth.py using root-cause-first analysis.
