Content
71%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.
The body is a well-structured, concise overview that delegates executable detail appropriately to a real, one-level-deep reference file. Its main weakness is actionability within the body itself — guidance names tools and directions but no runnable commands, and the workflow lacks an explicit feedback loop.
Suggestions
Inline one minimal executable command in the Check or Fix section (e.g., a one-line Lighthouse CLI invocation) so the body is immediately actionable without opening the reference.
Remove the opening rationale sentence and the Quick Reference bullets that restate why browsers matter — Claude already knows this, and the reference's 'Why It Matters' section covers it.
Add an explicit validation checkpoint after Fix (e.g., 'Re-run the audit and confirm the targeted metric improved before considering the change done') to close the workflow's feedback loop.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean (~30 lines) with organized sections, but the opening rationale sentence ('Static analysis alone cannot simulate the complex rendering, script execution, and layout processes of modern web browsers') and the Quick Reference bullets restate knowledge Claude already has and could be trimmed. | 4 / 5 |
Actionability | Specific tools (Lighthouse, Puppeteer, WebPageTest, DevTools) and metrics (LCP, CLS, INP) are named, and the Check/Fix sections give direction ('Ensure that performance audits are conducted in an environment that executes JavaScript and renders the page'), but the body contains no executable commands or code — the concrete examples are deferred to references/rule.md, leaving key details missing from the body itself. | 3 / 5 |
Workflow Clarity | A coherent Check → Fix → Explain → Code Review sequence is present with most checkpoints included ('Ensure that performance audits are conducted in an environment that executes JavaScript', 'describe the measurement method used to confirm the issue'), though an explicit validate-and-retry feedback loop is absent. | 4 / 5 |
Progressive Disclosure | A concise overview body clearly signals a one-level-deep reference ('For full implementation details, code examples, and framework-specific guidance, see references/rule.md'); the referenced file exists, contains substantial executable Lighthouse/Puppeteer content matching the pointer, and has no nested references. | 5 / 5 |
Total | 16 / 20 Passed |