Content
70%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is well-structured with a clear validated workflow and good sectioning, but actionability is held back by speculative/placeholder commands (the --raw flag and unexplained sed) and conciseness by repeated restatements of the source-of-truth rule.
Suggestions
Replace the speculative '# whatever flag exposes raw body' gbrain invocation with the verified, executable flag and a fallback, so the core path is fully copy-paste ready.
Add a one-line comment explaining the sed frontmatter-strip expression so it is unambiguous and reproducible.
Consolidate the source-of-truth rule to a single canonical statement to trim repeated restatements across the callout, 'The rule', 'Delivery', and 'Anti-Patterns' sections.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient with concrete code and tight rationale, but it repeats the source-of-truth rule across the convention callout, 'The rule', 'Delivery', and 'Anti-Patterns', and includes some redundancy that could be tightened, matching the score-2 anchor. | 2 / 3 |
Actionability | It provides mostly executable bash, but key steps are speculative or incomplete — "gbrain get "$SLUG" --raw > "$RAW" # whatever flag exposes raw body" and the unexplained frontmatter-strip sed — so it is not fully copy-paste ready, matching score-2. | 2 / 3 |
Workflow Clarity | The RESOLVE→STRIP→RENDER→DELIVER workflow is clearly sequenced with an explicit validation checkpoint (confirming the brain page and make-pdf binary exist before proceeding) and an anti-patterns list, matching the score-3 anchor. | 3 / 3 |
Progressive Disclosure | For a single-file skill under ~190 lines with no bundle files, content is well-organized into clearly signaled sections with one-level-deep links to conventions/related skills, satisfying the simple-skill scoring note for a 3. | 3 / 3 |
Total | 10 / 12 Passed |