Content
81%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
An exceptionally actionable QA workflow with executable code, explicit validation gates, and a genuine iteration loop, organized into clear sections with a real one-level-deep reference. Its only notable weakness is repetition and duplication between the inline fidelity-surface checklists and references/qa-rubric.md, which costs tokens without adding guidance.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with prescriptive, operational directives and explains nothing Claude already knows, but it repeats key instructions verbatim ('Put the source image and the implementation screenshot together in the same comparison input' appears twice) and the Required Fidelity Surfaces section duplicates checklists already in references/qa-rubric.md. | 3 / 5 |
Actionability | Fully executable throughout: a copy-paste-ready Playwright snippet with a 393 x 852 bounding-box guard, concrete commands ('npm run check:runtime'), specific selectors ('data-testid="device-screen"', '[data-phone-screen]'), a severity taxonomy, an output template, and an explicit design-qa.md field list. | 5 / 5 |
Workflow Clarity | A clear six-step sequence with explicit validation checkpoints (blocking 'npm run check:runtime', blocked-on-missing-artifacts rules), a documented find -> fix -> re-capture -> re-compare feedback loop, and a required-fields checklist for the final report. | 5 / 5 |
Progressive Disclosure | Good structure with a real, clearly signaled, one-level-deep reference ('Read [qa-rubric](./references/qa-rubric.md) when the QA pass spans more than a quick visual check'), but the inlined Required Fidelity Surfaces section substantially overlaps that reference file, leaving a minor content-splitting gap versus anchor 5. | 4 / 5 |
Total | 17 / 20 Passed |