Content
86%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 exceptionally lean, disciplined instruction skill: concrete tool contracts, explicit error-handling fallbacks, honest-sourcing requirements, and clean structure with no wasted tokens. The gaps are definitional — the remediation-type vocabulary and the expected output/placeholder formats are assumed rather than shown — which is what keeps actionability and workflow clarity at very good rather than excellent.
Suggestions
Define the remediation types once, e.g. a short line clarifying that the tool returns guidance (text-only), hybrid (text + commands), or executable (script) remediations, since behavior branches on these values.
Add one compact example of a substituted output line (resourceId-derived subscription, resource group, and resource names filled in) and one of a preserved unresolved placeholder so the substitution rule is unambiguous.
State how to retrieve Microsoft Learn documentation (e.g. searching learn.microsoft.com by recommendationTypeId or description) so the fallback path is executable rather than directional.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Every line is imperative instruction with zero padding and no explanation of concepts Claude already knows (no Azure primers, no tool-architecture background). The skill assumes competence and each sentence earns its place, matching the lean-and-efficient top anchor; the terse compressed phrasing is density, not bloat. | 5 / 5 |
Actionability | Concrete, executable guidance throughout: named tool calls with exact parameters ("advisor_remediation_get" with "recommendationTypeId", "advisor_recommendation_list"), a named error path ("RemediationNotFound"), exact output strings ("Guidance source: Azure Advisor remediation tool"), and CLI/PowerShell separation. Falls short of fully-executable because the remediation types ("guidance", "hybrid", "executable") are used without being defined and no example of the placeholder-substituted output format is given. | 4 / 5 |
Workflow Clarity | A clearly sequenced 5-step workflow with an explicit error-recovery branch (RemediationNotFound/empty → state unavailability → fall back to Learn documentation → ask or admit absence) and a final-response labeling requirement acting as an output checklist. Not the top level because the remediation-type terminology that gates command generation is never defined, and the placeholder substitution format is unspecified. This is not a destructive or batch skill — execution is explicitly forbidden without confirmation — so the validation-cap rule does not apply. | 4 / 5 |
Progressive Disclosure | The skill is ~30 lines, self-contained, and needs no external references; sections (Quick reference, Workflow, Final response) are well-organized and easy to navigate. Per the simple-skill guideline, this earns the top score on body organization alone. | 5 / 5 |
Total | 18 / 20 Passed |