Content
53%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 lean and well-organized with a clear niche and sensible operating principles, plus typos ('node_modeules', 'Guidlines') and Role/principles duplication that slightly undercut polish. Its main weakness is actionability: it tells Claude what to value but gives no concrete tokens, primitives, or examples, deferring almost everything to external links. A short decision workflow connecting the trigger list to the external resources would raise workflow clarity.
Suggestions
Add concrete, actionable specifics inline: 3-5 semantic token names, the primitive names for common cases (table, form field, navigation), and one good/bad copy example, so the skill instructs without requiring a web fetch.
Consolidate the Role bullets with Operating principles 1-2 to remove duplication, and fix the typos ('node_modeules', 'Guidlines') that undermine trust in the paths.
Add a short numbered consultation workflow (e.g. 1. identify the UI need, 2. check Storybook/docs for a matching primitive, 3. map values to semantic tokens, 4. verify accessibility states) so the numbered principles become an executable sequence.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is short and decision-level, avoiding explanations of concepts Claude already knows. Minor trimming is possible: the Role bullets ('choose the right primitive first', 'apply semantic tokens...') substantially duplicate Operating principles 1 and 2, and the 'latest version' note is filler. Fits anchor 4 (efficient, minor over-explanation) rather than anchor 5 (every token earns its place). | 4 / 5 |
Actionability | Guidance is high-level ('Use Aura primitives before custom UI', 'no raw values') with no concrete token names, component names, code snippets, or copy patterns; the only concrete artifacts are pointers to external docs. This matches anchor 2 (high-level hints missing the specific steps to execute) — it is above anchor 1 because the resource pointers (DESIGN.md path, Storybook URL) do give executable entry points. | 2 / 5 |
Workflow Clarity | Operating principles are numbered and principle 3 embeds a small decision loop ('check variants/props first, then document the gap'), but there is no sequence for an actual consultation — how to move from the when-to-reference triggers to picking a primitive, applying tokens, and verifying against the Storybook/docs. Fits anchor 3 (sequence present but checkpoints missing or implicit), and the operations are non-destructive so no cap applies. | 3 / 5 |
Progressive Disclosure | The body is under 50 lines with clear sections (Role, when-to-reference, Operating principles) and all references (node_modules/@cognite/aura, DESIGN.md, Storybook, docs site) are external and one level deep — good structure with minor gaps. Not anchor 5 because the resource list is buried mid-prose under 'Everything you need to know is:' rather than clearly signaled sections, and the local DESIGN.md reference is a path mention rather than an organized reference. | 4 / 5 |
Total | 13 / 20 Passed |