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.
An exemplary lean, instruction-only skill: a clear three-phase workflow with a tool fallback, concrete analysis heuristics Claude would not derive on its own, and disciplined rules on separating proven from tested creative. The only meaningful gap is that the ad-library fallback and analysis categories stay generic where named sources or a worked example would make execution fully turnkey.
Suggestions
Name specific public ad libraries in step 1 (e.g., Meta Ad Library, Google Ads Transparency Center) so the fallback is executable without guessing where to look.
Add a brief worked example of one analyzed ad (angle, offer, longevity, inferred move) to anchor the framework in a concrete case.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Every line carries non-obvious domain guidance — the tool name, "longevity alone does not establish profitability", "absence is opportunity", the trademark risk of copying — with zero concept explanation or padding. This matches anchor 5 (lean, assumes Claude's competence, every token earns its place); anchor 4 would require trimmable over-explanation, of which there is none. | 5 / 5 |
Actionability | Concrete guidance throughout: the named tool `research_campaign_opportunity`, an explicit fallback ("Ask for public ad-library links or examples when the tool does not return competitor ads"), specific analysis dimensions ("price framing, trial, guarantee, urgency"), and handoff to named skills ("ad-copy-generation or **launch** skill"). It stops short of anchor 5 because the fallback names no specific ad libraries and no worked example covers a common case, while clearly exceeding anchor 3's pseudocode-level guidance. | 4 / 5 |
Workflow Clarity | The three numbered phases (find what they run → read the pattern → turn it into your move) form a clear sequence, with an error-recovery checkpoint in step 1 and a validation rule ("Separate what is proven... from what is a test") in Rules. The skill is read-only so the destructive/batch validation cap does not apply, and the simple-skill exception holds for this under-50-line single-task skill with an unambiguous process — matching anchor 5 rather than anchor 4's 'minor validation gaps'. | 5 / 5 |
Progressive Disclosure | The skill is under 50 lines with no bundle files (no references/, scripts/, or assets/ exist) and no content that warrants external files; it is organized into four well-labeled sections (## 1, ## 2, ## 3, ## Rules) and only points to other skills, not nested documents. Per the rubric's guideline for short skills needing no external references, this matches anchor 5. | 5 / 5 |
Total | 19 / 20 Passed |