Content
73%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 operationally strong: a well-sequenced 9-step audit workflow with explicit validation discipline, decision gates, and copy-paste-ready commands and thresholds. Its weaknesses are moderate token padding (duplicated overview sections, re-explained JS-rendering basics, a strikethrough formatting glitch) and several references that resolve outside the provided bundle.
Suggestions
Cut the opening sentence and the 'What This Skill Does' section (both restate the frontmatter description), and trim the Audit Notes SSR/CSR primer to just the actionable instruction: compare raw fetch vs rendered DOM.
Fix the broken strikethrough in Data Sources ('Use ~~web crawler, ~~page speed tool, and ~~CDN') — the stray tildes render as struck-through text and obscure which connectors are meant.
Consolidate the long Data Sources paragraphs (crt.sh recipe, W3C validator, IndexNow/Baidu push details) into a reference file like the other playbooks, keeping only the zero-dependency helper list inline.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient but padded in places: the opening line and the 'What This Skill Does' section restate the frontmatter description, and Audit Notes re-explains JS-rendering basics Claude already knows ('SSR/SSG ships it server-side; pure CSR hides it until hydration'). A mangled '~~web crawler, ~~page speed tool, and ~~CDN' strikethrough glitch adds noise; anchor 4 would require only minor trims. | 3 / 5 |
Actionability | Concrete, executable-shaped guidance throughout — `python3 "${CLAUDE_PLUGIN_ROOT}/scripts/connectors/robots.py" <url> --check-ai-bots`, the crt.sh curl recipe, the ledger-diff fix-verification pipeline, and explicit CWV thresholds (LCP <2.5s, INP <200ms, CLS <0.1). Scored 4 rather than 5 because the connector scripts referenced via ${CLAUDE_PLUGIN_ROOT} are not present in the provided bundle, leaving the commands unverifiable here. | 4 / 5 |
Workflow Clarity | A numbered 9-step audit sequence with a per-step evidence requirement, metric provenance labeling ('Measured', 'User-provided', 'Estimated', N/A rule), explicit Decision Gates (stop-and-ask vs. continue-silently), done-when criteria, and a re-run ledger-diff feedback loop to prove fixes — clear sequence with explicit validation and error-recovery checkpoints. The bulk-audit batch path is explicitly gated, so the batch-cap does not apply. | 5 / 5 |
Progressive Disclosure | Eight one-level-deep reference files, each clearly signaled inline and indexed with descriptions under Reference Materials; however, four paths escape the skill bundle (../../../references/skill-contract.md, ../../../CONNECTORS.md, ../../../SECURITY.md, ../../../scripts/connectors/README.md) and are not resolvable within it, and the dense Data Sources section could partly move to a reference file. Good structure with minor organization gaps rather than the fully clean split of anchor 5. | 4 / 5 |
Total | 16 / 20 Passed |