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.
The body is an unusually strong operational skill: exact commands, a real validation gate before any output is produced, fail-open handling of missing inputs, and explicit error verdicts. Its weaknesses are length and layout — rhetorical passages that could be one-line rules, and engine tables plus a report template that belong in reference files — which hold conciseness and progressive disclosure at 3.
Suggestions
Trim the "observed in practice" failure-mode anecdote and the "absence of evidence" paragraph to single-line rules — the directives stand without the diagnostic history.
Move the engine-specific scanning sections (Unreal Content/ and Unity Assets/ adaptations plus the per-content-type glob tables) into a references/engine-scanning.md, keeping only the engine-detection decision inline in SKILL.md.
Consider moving the full report template to references/report-template.md and inlining only the required fields and the verdict rules.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The bulk is dense operational instruction, but several passages could be tightened: the ~10-line "failure mode this prevents is specific and has been observed in practice" anecdote ("an asset audit returning COMPLIANT on a project with no assets... a performance profile reporting >99% headroom"), the "Absence of evidence is never evidence of absence" philosophical paragraph, and a tier-resolution note repeated between the top blockquote and the workflow section. Not a 2 (most content is actionable instruction, not padded explanation); not a 4 (the rhetorical/diagnostic history exceeds minor trims). | 3 / 5 |
Actionability | Fully executable guidance throughout: copy-paste-ready Grep invocations with exact regexes ("^## Summary"; the digit-plus-content-noun pattern in scan (b)), exact glob patterns per engine type, precise status thresholds ("IN PROGRESS — Found is 50–99% of Specified"), and a complete fill-in report template covering the common output case. Common engine cases (Godot, Unity, Unreal) are each explicitly handled. | 5 / 5 |
Workflow Clarity | Phases 1–5 are explicitly sequenced with a genuine pre-flight validation checkpoint ("record FOUND or ABSENT — not 'assumed present'"; stop conditions before any report is produced), fail-open rules for missing inputs (a GDD without ## Summary is full-read, never dropped), explicit error states with recovery routing ("NOT ASSESSED — NO DATA" naming which skill produces the missing input; "NOT ASSESSED — code root unresolved" instead of a false zero), and a closing verdict enum that prevents false COMPLETE results. This is the anchor-5 pattern: clear sequence, explicit validation, error-recovery loops. | 5 / 5 |
Progressive Disclosure | No bundle files exist (references/, scripts/, assets/ are all absent), so all ~310 lines live inline in SKILL.md. Section headers are clear and consistent, but content that would naturally sit in a reference file — the engine-specific Unreal/Unity scanning tables and the full report template — is inlined, and the only file references are paths into project docs (.claude/docs/workflow-modes.md, design/gdd/*.md) rather than a one-level-deep skill bundle. Matches "some structure but content that should be separate is inline"; not a 2 since structure and navigation within the file are good. | 3 / 5 |
Total | 16 / 20 Passed |