Content
92%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-engineered reference-skill body: index-first progressive disclosure, concrete MUST rules, literal response phrasings for gated/stub content, explicit error branches, and a freshness procedure. The only weakness is minor duplication of the events/showcase stub guidance and some overlap between the routing table and the section prose.
Suggestions
Deduplicate the events/showcase stub guidance: state the 'deliberately minimal stubs → hand over live URL' rationale once (Live-URL redirects) and have Scope notes reference it in one line instead of restating it.
Fold the routing-hints table's pointer rows (events, showcase) into a single 'stubs → live URL' row and drop rows that merely restate the Reference pages table, tightening the body without losing navigation.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and operational (routing table, MUST rules, keyword-filter procedure) with essentially no explanation of concepts Claude already knows. It sits at 'efficient; minor instances of over-explanation that could be trimmed' rather than 5 because the events/showcase stub rationale is stated twice ("Live-URL redirects" section and "Scope notes" — e.g. "The events and showcase pages are deliberately minimal stubs" repeats the earlier "Two pages are deliberately minimal stubs"), and the routing table re-summaries mappings the section prose already covers. | 4 / 5 |
Actionability | Guidance is copy-paste ready for an instruction-only skill: exact file paths ("references/learn.md", "references/faq.md"), a concrete procedure ("filter by keyword first, then open the linked detail file"), literal response phrasings ("respond along the lines of: 'Virtual and in-person AWS Startups events happening around you can be found at https://aws.amazon.com/startups/events.'"), hard rules ("MUST quote verbatim", "MUST cite source_url"), and an exact command ("npx skills update knowledge-base-for-startups"). Specific examples cover the common query cases via the routing table — a clear score-5 match. | 5 / 5 |
Workflow Clarity | The lookup workflow is clearly sequenced: open the index first, "filter by keyword, then open at most one detail file per question", answer with verbatim quote + source_url, then the post-answer freshness check with an explicit condition ("If the gap is more than 6 months"). Error-recovery branches are explicit: "If the index points to a sign-in-gated row, do not attempt to open the detail file — surface the title + Live URL" and "If the topic is not covered, say so plainly rather than improvising". No destructive/batch operations exist, so no validation cap applies; this matches the score-5 anchor rather than score 4's 'minor validation gaps'. | 5 / 5 |
Progressive Disclosure | SKILL.md is a pure routing overview pointing to well-signaled index files (learn.md, offers.md, build.md) which in turn link one level into the detail trees; every referenced path was verified to exist in the bundle (references/learn/<category>/<slug>.md, references/offers/<slug>.md, all reference pages and stubs). Each reference is described with its purpose and navigation pattern ("They have a Keywords column — filter by keyword first"), matching the 'clear overview with well-signaled one-level-deep references; easy navigation' anchor. | 5 / 5 |
Total | 19 / 20 Passed |