Flaky, skipped, and generated tests¶

R-SKIP-1 - Skipped tests have owner and reason¶

Rule. Skipped, todo, quarantined, or disabled tests must include a reason, owner, and review trigger.

Why. Skips without accountability become permanent blind spots while the test suite still reports green.

vcqa. Flag .skip, .todo, quarantine lists, disabled specs, or excluded test paths without nearby owner/reason metadata.

References.

• Vitest Guide: https://vitest.dev/guide/
• Playwright Best Practices: https://playwright.dev/docs/best-practices

R-FLAKE-1 - Flakes are fixed or quarantined explicitly¶

Rule. Flaky tests must be fixed promptly or quarantined with a bounded policy; retries alone are not a fix.

Why. Flakes teach teams to ignore CI and can hide real regressions behind retry noise.

vcqa. Flag high retry counts, repeated failure annotations, or flaky-test comments with no linked issue, owner, or expiry.

References.

• Playwright Best Practices: https://playwright.dev/docs/best-practices

R-GEN-1 - Generated tests must prove behavior¶

Rule. AI-generated or mechanically generated tests must meet the same assertion, fixture, and boundary-quality rules as human-authored tests.

Why. Generated tests often inflate coverage while asserting only mocks, existence, or tautologies.

vcqa. Flag bursts of new tests with weak assertions, excessive mocks, identical test shape, missing failure paths, or names that do not match the behavior asserted.

References.

• Testing Library Guiding Principles: https://testing-library.com/docs/guiding-principles