Content
93%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 review-lens skill body: dense, expert-level, and immediately actionable, with a clear what-to-check / how-to-judge / what-to-report structure and explicit exclusions preventing noise. The only improvement would be an explicit verification step in the workflow, though for a code-review skill the current checkpoints are adequate.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and expert-level: every sentence carries specific diagnostic content (exact property wrappers, exact APIs like 'performAndWait', 'NSManagedObjectID', 'DispatchSemaphore.wait()', exact build settings like 'OTHER_SWIFT_FLAGS'), with zero explanation of concepts Claude already knows and no padding. Nothing to trim. | 5 / 5 |
Actionability | Though it is an instruction-only skill with no code blocks, the guidance is fully concrete: it names exact wrong/right API pairs ('.sink { self.value = $0 } ... without [weak self]', '@ObservedObject on an object the view creates (should be @StateObject)'), the exact checks to run ('check whether the project builds with -strict-concurrency=complete or Swift 6'), and the exact fixes ('[weak self] or a stored cancellable', '@MainActor or perform', 'Decimal'). Per the rubric's scoring note, absence of code in an actionable instruction-only skill is not penalized. | 5 / 5 |
Workflow Clarity | The Method → Threshold → Reporting sections form a coherent review sequence with decision checkpoints (verify strict-concurrency/Swift version before judging severity; read the AGENTS.md/CLAUDE.md chain for project rules), but there are no explicit validation or error-recovery steps — matching the 4 anchor ('clear sequence with most checkpoints present'). Not 5 because no explicit verify/feedback loop exists; not 3 because the sequence is well-defined and this is a non-destructive review task where the cap does not apply. | 4 / 5 |
Progressive Disclosure | The body is under 50 lines, needs no external reference files (none exist in the bundle), and is organized into clearly labeled sections (Scope, Method, Threshold, Reporting) — meeting the rubric's simple-skill guideline for a top score. | 5 / 5 |
Total | 19 / 20 Passed |