Content
43%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 a clean, token-efficient catalogue stub: well-organized sections, no padding, no over-explanation. But as a skill it delegates nearly everything — the actual design workflow, scripts, and references — to an upstream GitHub URL with no local bundle files and no concrete install steps, leaving both actionability and workflow clarity at anchor 2.
Suggestions
Vendor the essential upstream material locally (or at minimum the core design workflow in SKILL.md) so the skill is actionable without an external round-trip to a README.
Make the install step concrete and executable: give the actual clone/copy command and target skills-directory path instead of 'install the upstream bundle' plus 'open' the README for 'exact paths'.
Add a validation checkpoint after install (e.g., verify the skill directory exists / the agent lists `canvas-design` as available) before instructing the agent to invoke it.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean (~28 lines), well under any token-budget concern, explains nothing Claude already knows, and never pads. It falls short of anchor 5 ('every token earns its place') due to verbatim duplication: the 'What it does' section repeats the frontmatter description word-for-word, and the upstream URL appears in both frontmatter and the 'Source' section — minor redundancy that keeps it at anchor 4 ('efficient; minor instances that could be trimmed'). | 4 / 5 |
Actionability | The only executable command is `open https://github.com/...` with the comment '# Inspect the upstream README for exact paths', and the core instruction — 'install the upstream bundle into your active agent's skills directory' — is a high-level hint with no install command, no paths, and no concrete steps. This matches anchor 2 ('minimal concrete guidance; high-level hints but missing the specific steps to execute') rather than 3, because there is no partially-complete executable guidance for the skill's actual purpose, only delegation to an external README. | 2 / 5 |
Workflow Clarity | A rough sequence exists (open upstream URL → install bundle → invoke by name or trigger), but each step is poorly defined: no installation command, no target paths, no confirmation the bundle installed correctly, and no validation before invocation. This is anchor 2 ('rough sequence present but many gaps; steps poorly defined; validation absent'); anchor 3 would require steps that are at least concretely defined even if checkpoints are missing. | 2 / 5 |
Progressive Disclosure | Sections are clearly organized ('What it does', 'Source', 'How to use') and the upstream reference is explicitly signaled, but the bundle contains no local reference files at all (no references/, scripts/, or assets/ directories) — all substantive material is deferred one hop away to an external GitHub repo. This does not qualify for the sub-50-line 5-exception (which requires no need for external references), and the single pointer to an out-of-band upstream README matches anchor 3 ('some structure; references present but not part of a navigable bundle') rather than 4's 'most content is appropriately placed'. | 3 / 5 |
Total | 11 / 20 Passed |