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.
A well-structured, actionable skill body with clear sequenced workflows, explicit validation steps, and exemplary progressive disclosure to a verified one-level-deep reference. Its main weakness is redundancy: the silent-ignore fact and the validation steps are each repeated across two or three sections, and the Code Review section largely duplicates Check.
Suggestions
Remove the duplication of the silent-ignore fact: it appears in the intro, Quick Reference, and Explain — state it once (ideally in the intro) and cut the rest.
Merge or cut the Code Review section, which restates the Check steps (parse, verify @context, check @type, validate required properties) almost verbatim.
Tighten the Explain section: what rich results are and that Google uses structured data are concepts Claude already knows; keep only the non-obvious insight (validation failures are only discoverable via the Rich Results Test).
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The bulk (Check, Fix) is efficient and actionable, but the silent-ignore fact is stated three times (intro, "Invalid JSON (unclosed brackets, missing commas) causes the entire block to be silently ignored", and the Explain section), and the Code Review section restates the Check steps almost verbatim. This fits "mostly efficient but includes some unnecessary explanation or could be tightened" rather than anchor 2, since there is no concept-explanation padding and over half the body is tight actionable content. | 3 / 5 |
Actionability | Concrete, executable guidance throughout: "Parse the JSON-LD with JSON.parse()", the exact "https://schema.org" value, enumerated required properties per type (Article: headline, author, datePublished; Product: name; FAQPage: mainEntity; BreadcrumbList: itemListElement), and two validation URLs. Minor gaps keep it below anchor 5: no inline code examples (delegated to the reference, acceptable for an instruction-only skill) and "Run through Google's Rich Results Test API if available" lacks an endpoint or command. | 4 / 5 |
Workflow Clarity | Check (5 numbered steps) and Fix (6 numbered steps) are clearly sequenced, each terminating in explicit validation steps (Rich Results Test, jsonlint). It falls short of anchor 5 because no explicit feedback loop ("if validation fails, fix and re-validate") connects the sections, though this is a non-destructive single-page task so no cap applies. | 4 / 5 |
Progressive Disclosure | The body is under 50 lines, cleanly sectioned (Quick Reference, Check, Fix, Explain, Code Review), and closes with a clearly signaled one-level-deep reference — "For full implementation details, code examples, and framework-specific guidance, see `references/rule.md`" — which exists as a real 160-line file with code examples. The split is appropriate: overview and core workflow inline, full examples in the reference. | 5 / 5 |
Total | 16 / 20 Passed |