Content
75%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-crafted, lean instruction-only skill: concrete per-platform rules, literal banned-phrase examples, an untrusted-content section, and a quality-gate checklist give it strong actionability and clear workflow. Its residual gaps are minor — a small amount of rule repetition, no worked example of an adapted variant, and workflow steps that assume external skills (brand-voice, content-engine) without fallback guidance.
Suggestions
Add one compact worked example (source line → X variant → LinkedIn variant) to the Output Format section; concrete before/after examples would push actionability toward anchor 5.
State what 'brand-voice' and 'content-engine' return (or the fallback if unavailable) at the workflow steps that invoke them, closing the workflow_clarity gap — e.g. 'if brand-voice is not present, derive a minimal voice profile from the source directly'.
Dedupe the identical-copy rule: it appears as Core Rule 1, in the frontmatter description, and again in the Quality Gate — keeping it in the Core Rules and the Quality Gate alone would trim tokens without losing the constraint.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and imperative throughout — short bullets, no concept explanations Claude already knows, no filler ('keep it compressed', 'do not write fake hyper-casual creator copy'). Minor redundancy keeps it from anchor 5: 'Do not publish identical copy across platforms' appears as Core Rule 1 and is restated in the Quality Gate ('no copy is duplicated verbatim across platforms') and in the frontmatter description. Well above anchor 3 — there is no unnecessary explanation to trim, just repetition. | 4 / 5 |
Actionability | Concrete, executable instruction-only guidance: literal banned phrases to delete ('Excited to share', 'Here's what I learned', 'link in bio'), specific per-platform do/don't rules, a defined output format, and an explicit workflow. Per the rubric's scoring notes, absence of code is not penalized for instruction-only skills. It falls short of anchor 5 only because there is no worked example of an adapted variant or a note on what changed — the guidance covers the rules but not a concrete before/after case. | 4 / 5 |
Workflow Clarity | A clear three-step sequence (pick primary version → capture voice fingerprint via brand-voice → adapt by platform) followed by posting order and an explicit pre-delivery validation checklist ('Before delivering: each version reads like the same author...'). This is not a destructive or batch operation, so the validation cap does not apply. Anchor 4 rather than 5 because the middle steps depend on external skills (brand-voice, content-engine) without stating what they return or what to do if they are unavailable — a minor checkpoint gap. | 4 / 5 |
Progressive Disclosure | A single, self-contained SKILL.md (~110 lines) with well-organized sections, no nested references, and related skills clearly signaled in a dedicated 'Related Skills' section. No bundle files exist (no references/, scripts/, or assets/ directories), so all content lives appropriately in one file. Anchor 4 rather than 5: the body exceeds the under-50-lines simple-skill exception, and the per-platform constraint blocks plus Banned Patterns are candidates that could split into a reference file if the skill grows. | 4 / 5 |
Total | 16 / 20 Passed |