Content
81%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 tightly written, highly actionable analysis workflow: exact formulas, hard numeric thresholds, a gating validation phase with a fallback mode, and a mandatory report structure, all in third-person imperative voice with explicit output standards and scope exclusions. The only improvements are minor — slight intro redundancy, one vague data-pull instruction, and templates that could be split into reference files.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense, table-driven content with no tutorial padding — it states its own thesis ("The lowest-CPA campaign is often the one bringing in the *worst* leads") rather than explaining concepts Claude already knows — but the "Core principle" paragraph partially restates the intro and the "When to Use" list duplicates the frontmatter description, so minor trims remain. | 4 / 5 |
Actionability | Concrete executable guidance throughout: exact formulas (True CAC = spend ÷ qualified leads), hard thresholds (≥80%/50–80%/<50% joinable, ≥30 signups, 14-day window), a sample size (10–20 signups), classification rules per bucket, and an exact 7-section report template; the only gap is that pulling Meta-side data ("via the existing Meta Marketing API connection") lacks an example call or query. | 4 / 5 |
Workflow Clarity | Phases 0–6 are clearly sequenced with explicit validation checkpoints and feedback loops: a discovery gate ("Don't proceed until each is answered"), a gating tracking-validation step with coverage thresholds, a hard fallback (<50% joinable → tracking-gap mode, skip Phases 2–6), cohort-maturity guards on Cut decisions, and confidence flags on every recommendation. | 5 / 5 |
Progressive Disclosure | No bundle files exist and the skill is self-contained with clear, well-ordered sections and no buried references; the report and tracking-gap output templates are inlined where they could plausibly live in separate reference files, which is a minor organization gap rather than a structural problem. | 4 / 5 |
Total | 17 / 20 Passed |