Content
67%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 well-structured, highly actionable instruction skill with a clear workflow, explicit validation checkpoints, and a disciplined reference-pointing design. The main costs are a large hidden metadata comment block and overlapping trigger sections that inflate the token budget, the absence of a concrete example hook entry, and reference files that are named but not actually present in the bundle.
Suggestions
Remove or slim the hidden HTML CAPABILITIES_SUMMARY / COLLABORATION_PATTERNS comment block — its content is already duplicated by the visible Collaboration and Trigger Guidance sections and costs ~40 lines of context every load.
Add one minimal, copy-paste-ready example of a settings.json PreToolUse matcher-group hook entry (with timeout and exit-2 blocking pattern) so the common 'configure' case is fully executable from the body.
Consolidate the 19-bullet Trigger Guidance list with the Signal Keywords table — they overlap heavily and could be one dispatch section, cutting redundant tokens.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense, table-driven, and almost entirely domain-specific rather than explaining known concepts, but it carries real padding: the ~40-line hidden HTML comment block (CAPABILITIES_SUMMARY, COLLABORATION_PATTERNS, BIDIRECTIONAL_PARTNERS) duplicates the body's own Collaboration and Trigger Guidance sections, and the 19-bullet Trigger Guidance overlaps the Signal Keywords section. That is more than the 'minor instances' of a 4 but not the pervasive over-explanation of a 2. | 3 / 5 |
Actionability | Concrete commands ("jq . ~/.claude/settings.json", "/hooks", "claude --debug"), exact event names, matcher patterns, exit codes, timeouts, and a subcommand dispatch table make the guidance mostly executable. It falls short of anchor 5 because there is no copy-paste-ready example of a settings.json hook entry or hook script to cover the common configure case. | 4 / 5 |
Workflow Clarity | SCAN → PROPOSE → IMPLEMENT → VERIFY → MAINTAIN is clearly sequenced with per-step goals and reads, and validation checkpoints are explicit (backup before modify, jq validation, the VERIFY step, the Always checklist). Not a 5 because the inline error-recovery loop ('if validation fails, fix and re-validate') is delegated to references rather than stated in the workflow. | 4 / 5 |
Progressive Disclosure | Structure is good: a Reference Map with 'read this when' guidance, per-recipe 'Read First' columns, and per-workflow-step reads signal one-level-deep references clearly. It cannot score 5 because none of the referenced files (reference/hook-system.md, reference/event-catalog.md, _common/*.md) exist in the bundle, and some contract detail that the references should hold is inlined in the body; yet the in-body organization is clearly above the buried-reference anchor 3. | 4 / 5 |
Total | 15 / 20 Passed |