| name | quality-assurance-guidelines |
| description | Use this skill when reviewing whether a change has adequate verification evidence. Covers format/lint/build/test proof, unit coverage for password and layout logic, e2e coverage for generator flows, clipboard and responsive manual checks, flaky-test handling, skipped commands, residual risk, and second-pass verification after fixes. |
Quality Assurance Guidelines
Apply these rules when judging whether a change has been adequately verified before completion or merge. This is the reviewer's QA lens on top of Development Guidelines and E2E Testing Guidelines.
Verification Evidence
See verification-evidence.md for how to report commands, manual checks, skipped checks, and residual risk.
Guidelines:
- SHOULD read the linked reference when reviewing completion evidence or preparing final verification reporting.
Lint and Format Gate
See lint-and-format-gate.md for Biome format/lint expectations, biome-ignore review, and touched-file warning handling.
Guidelines:
- SHOULD read the linked reference when work edits code, docs, skills, config, tests, or generated-word scripts.
Unit and E2E Coverage
See test-coverage.md for expected coverage by surface: password generation, keyboard layouts, React UI, routing, and deployment config.
Guidelines:
- SHOULD read the linked reference when reviewing test adequacy or deciding whether new tests are required.
Manual Verification
See manual-verification.md for browser checks that automated tests do not fully cover: responsive layout, focus, copy feedback, and failure states.
Guidelines:
- SHOULD read the linked reference when UI output, browser APIs, or responsive behavior changes.
Flakiness Tolerance
See flakiness-tolerance.md for handling intermittent Playwright or Vitest failures without weakening coverage.
Guidelines:
- SHOULD read the linked reference when a test is retried, skipped, marked flaky, or changed to wait for asynchronous browser behavior.