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.
Highly actionable content: exact commands, a repository map, validated install flows (dry-run first, ask before apply), and ready-to-use output templates. It is slightly held back by a redundant 'When To Use' section duplicating the frontmatter and the absence of explicit error-recovery loops in troubleshooting.
Suggestions
Trim or collapse the 'When To Use' bullet list, which repeats the frontmatter description's triggers almost verbatim, into a single line to cut duplicate tokens.
Add an explicit error-recovery step to Troubleshooting (e.g., what to inspect when 'npm run harness:audit' reports failures and when to re-run it) to close the feedback-loop gap.
Consider moving the Output Templates section into a references/ file (e.g., templates.md) and linking to it, keeping the main body as a lean overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude's competence — no background concept explanations, just commands, paths, and templates. The 'When To Use' section (lines 12-22) largely restates the frontmatter description's triggers, which is duplicative padding that could be trimmed, keeping it just below anchor 5. | 4 / 5 |
Actionability | Fully executable, copy-paste-ready commands throughout: 'node scripts/ci/catalog.js --json', 'find skills -maxdepth 2 -name SKILL.md | sort', 'node scripts/install-apply.js --profile minimal --target claude --dry-run', 'rg -n "<query>" skills commands agents docs', plus a repository map with exact paths and concrete output templates covering the common cases (recommendation, search results, install plan summary). | 5 / 5 |
Workflow Clarity | Common Tasks have clear sequences with real checkpoints — install flow mandates '--dry-run' before apply, project onboarding includes 'ask before applying changes' as step 4, and troubleshooting directs asking for target harness first. However, error-recovery feedback loops (what to do when a dry run or audit fails and how to retry) are implicit rather than explicit, so it sits between anchors 4 and 5. | 4 / 5 |
Progressive Disclosure | No bundle files exist (no references/, scripts/, or assets/ directories), so this is a single, well-organized SKILL.md with clear sections and only one-level-deep pointers to files in the ECC repo ('README.md', 'hooks/hooks.json', 'manifests/install-*.json') rather than nested references. Minor gap: at ~190 lines, content like the Output Templates section is borderline for being split into a reference file, so it does not fully match anchor 5's ideal split. | 4 / 5 |
Total | 17 / 20 Passed |