Content
77%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 highly actionable, well-sequenced read-only analysis procedure with concrete commands, report templates, and severity heuristics. Its weaknesses are token redundancy from a verbatim-duplicated extension-hook block and a monolithic single-file layout where reference files would serve progressive disclosure.
Suggestions
Deduplicate the extension-hook procedure: the 'Pre-Execution Checks' section and Execution Step 9 repeat ~40 lines each; define the hook protocol once and reference it (or move it to a hooks.md reference file covering both before_analyze and after_analyze).
Split detailed material into one-level-deep references (e.g., report-format.md for the table templates and metrics, detection.md for the six detection passes) so SKILL.md stays a lean overview.
Remove or relocate the shell-quoting aside about "I'm Groot" — it is unrelated to the analysis workflow and adds noise to the prerequisites step.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly tight operational instruction with no explanations of concepts Claude already knows, but it carries clear redundancy: the extension-hook procedure (parse extensions.yml, filter enabled, condition handling, dot-to-hyphen command mapping, optional/mandatory output templates) appears twice nearly verbatim in "Pre-Execution Checks" and "Execution Steps > 9" (~80 combined lines), plus a misplaced quoting-escape aside about "I'm Groot". This fits anchor 3 (mostly efficient but could be tightened) more than anchor 2, since the padding is duplication rather than concept over-explanation. | 3 / 5 |
Actionability | Fully executable guidance throughout: the exact prerequisite command (".specify/scripts/powershell/check-prerequisites.ps1 -Json -RequireTasks -IncludeTasks"), exact artifact path derivations, a concrete report template with an example findings row and metric list, named severity tiers with examples ("one requires Next.js while other specifies Vue"), and explicit abort instructions. Copy-paste-ready commands and templates cover the common cases, matching anchor 5. | 5 / 5 |
Workflow Clarity | Nine numbered execution steps are clearly sequenced, starting with a prerequisites gate (abort with instructions if files missing), followed by artifact loading, semantic model construction, six named detection passes, severity assignment, a structured report, severity-gated next actions, and opt-in remediation. Read-only operation is enforced throughout, zero-issue handling is specified, and deterministic reruns are required — matching anchor 5's explicit checkpoints and error-recovery guidance. | 5 / 5 |
Progressive Disclosure | No bundle files exist (references/, scripts/, assets/ are absent) and the skill is a single ~245-line SKILL.md with good section headers. However, the duplicated hook procedure (~80 lines) is content that clearly belongs in a shared reference file, and a file this length should split detailed material (e.g., hook protocol, report format) into one-level-deep references. This fits anchor 3 (some structure, content that should be separate is inline) better than anchor 4, where most content is appropriately placed across files. | 3 / 5 |
Total | 16 / 20 Passed |