Content
88%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 tight, well-structured instruction skill: lean token usage, an explicit hit-to-live-case verification loop, and a clear exit gate. The two weaknesses are minor: one pattern entry is descriptive rather than a literal regex, and two references (tester-static-rules.json, 'the matrix') resolve outside the skill's bundle.
Suggestions
Rewrite the fetch row as a literal, copy-paste-ready regex like the other rows (e.g. 'fetch\([^)]*\)\s*;?(?![\s\S]{0,200}\.ok)' or a grep command) so every pattern entry is executable as written.
The referenced 'tools/harness/tester-static-rules.json' is not part of this skill's bundle — either ship it in a references/ directory or state where it lives so the path is resolvable; do the same for 'the matrix' (name the file or sibling skill that defines it).
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~30-line body is lean and dense — a compact pattern table, one-line procedure steps, and an exit-criteria gate — with no explanation of concepts Claude already knows; every token earns its place, matching the top anchor. | 5 / 5 |
Actionability | Mostly executable guidance: literal regexes like 'innerHTML\s*=', 'catch\s*\(\w*\)\s*\{\s*\}', 'outline:\s*none', and 'sk_live_' plus a concrete fetch-run-record procedure. Minor gaps keep it below 5: the fetch row ('await fetch( with no .ok / status nearby before navigation') is descriptive rather than a literal regex, and 'write a live case ID in the matrix' references a matrix not defined in this skill. | 4 / 5 |
Workflow Clarity | A clear 3-step sequence (fetch served code, run patterns and record locations, convert each hit to a live case) with explicit validation via the exit criteria — 'Every hit is either converted into a live case or dismissed with a written reason' — and a built-in verification loop ('Each one must become a live case before it can be called a bug'). The scan is read-only, so the destructive/batch cap does not apply. | 5 / 5 |
Progressive Disclosure | Sections are well organized for a sub-50-line skill, but the 'Learned rules' section points to 'tools/harness/tester-static-rules.json', which does not exist in this skill's bundle (no references/ or scripts/ directory is present), and 'the matrix' similarly points outside the skill — 'references mostly clear; minor organization gaps' rather than fully navigable. | 4 / 5 |
Total | 18 / 20 Passed |