Content
82%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 exemplarily terse, high-signal atom: concrete tool commands, explicit edge-case rules, and a well-sequenced workflow with a conditional fallback mode. Its main gaps are heavy dependence on an external measurement contract for key mechanics and a reference layout that lives outside the skill's own bundle.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and efficient throughout: the body assumes competence, explains nothing Claude already knows, and packs concrete rules ("never totalled by eye", "de-duplicate by ad id", "no mid-run asks") into ~16 lines with no padding. | 5 / 5 |
Actionability | Mostly executable guidance: named tool reads (`meta_read performance`, `social_analytics_get`, `ads_creative_read`, `video_read`, `campaign_read`), a concrete save path (`reports/performance/inputs/<YYYY-MM-DD>.md`), and explicit decision rules. Not a 5 because key mechanics (minimum read, totalling rules) are deferred to 'contract §4' in an external file that cannot be verified from this bundle. | 4 / 5 |
Workflow Clarity | A clear numbered sequence (1–5, with 3b) covering intake, hygiene, maturity comparison, persistence and coverage, plus an explicit conditional branch to claims-only mode when the referee is missing. Checkpoints like the coverage line and 'mark the referee not measured until it arrives' are present but some validation is implicit rather than an explicit verify step, keeping it below anchor 5. | 4 / 5 |
Progressive Disclosure | Good structure: a clearly signaled one-level-deep reference up front ("Load: `../analyze-performance/references/measurement-contract.md`") with the body kept as a lean overview. Not a 5 because the reference points into a sibling skill's directory rather than a local bundle (no references/, scripts/ or assets/ exist here), so navigation and placement have a minor organizational gap. | 4 / 5 |
Total | 17 / 20 Passed |