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 exceptionally lean, high-signal skill body: exact tool names, parameters, error states, and verbatim output labels, with a gated workflow and an explicit fallback loop. The only weakness is the absence of a worked example (a sample tool invocation or a before/after placeholder substitution), which would make the guidance fully copy-paste ready.
Suggestions
Add one example `advisor_remediation_get` invocation showing the `recommendationTypeId` GUID format to make the tool call copy-paste ready.
Include a short before/after example of placeholder substitution from a `resourceId` into a remediation step, so the 'preserve and describe unresolved placeholders' rule is concrete.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~25-line body contains zero background exposition (no explanation of what Azure Advisor is) and no padding; every sentence is directive, e.g. "Call `advisor_remediation_get` once with only `recommendationTypeId`, never `resourceId`". It fully assumes Claude's competence, matching anchor 5. | 5 / 5 |
Actionability | Guidance names the exact tool (`advisor_remediation_get`), its exact parameter, the exact error state (`RemediationNotFound`), and verbatim response labels ("Guidance source: Azure Advisor remediation tool"). As an instruction-only skill this is actionable without code, but minor gaps remain: no example invocation showing the parameter format and no example of a substituted placeholder in a final response, so it sits at anchor 4 rather than 5. | 4 / 5 |
Workflow Clarity | The 5-step workflow has ordering gates ("before generating guidance", "After retrieval, substitute"), an explicit error-recovery loop (RemediationNotFound/empty → Microsoft Learn fallback → state unavailability, "never invent guidance"), and ambiguity checkpoints ("Ask which resource if ambiguous; never guess", "reuse known values, ask for missing values, never guess"). This matches anchor 5's explicit validation and feedback loops; anchor 4's 'minor validation gaps' does not apply. | 5 / 5 |
Progressive Disclosure | This is a single-purpose skill under 50 lines with no external reference material needed, and no references/scripts/assets bundle exists. Per the rubric's simple-skill guideline, well-organized sections (Quick reference / Workflow / Final response) merit anchor 5 — and the sections are indeed clear and non-overlapping. | 5 / 5 |
Total | 19 / 20 Passed |