Content
71%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 well-organized, highly concrete design reference with excellent numeric defaults and a clear output procedure. Its weaknesses are structural (a ~185-line single file with no progressive disclosure to reference files), minor restatement of animation-performance knowledge Claude already has, and a few high-level sections (morph/shared-element, hero timelines) without actionable recipes.
Suggestions
Split the detailed catalogs into reference files (e.g., references/patterns.md for choreography/hover/scroll specs, references/defaults.md for the duration/easing/stagger tables) and keep SKILL.md as a lean overview with clearly signaled one-level links — this addresses the lowest-scoring dimension, progressive_disclosure.
Trim restatements of standard knowledge (the transform/opacity property rules, reflow/GPU advice) to a single non-negotiable rule line each, keeping the body focused on the curated numbers Claude does not already know.
Make the 'Output format' section explicitly consume the answers from 'What to ask the user' (e.g., brand lane → easing set, key moments → choreography plan) so the workflow closes the loop, and add a morph/shared-element recipe to close the actionability gap.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is predominantly lean, opinionated bullet guidance with concrete numbers ("Micro (hover/press): 120–200ms", "40–90ms per element"), but it restates knowledge Claude already has ("Prefer: transform (translate/scale/rotate), opacity / Avoid: width/height/top/left", "Don't measure layout every frame") and carries light manifesto padding ("This skill helps you ship tasteful, product-grade motion. Not 'more animation.'"). That matches anchor 4 — efficient with minor trims available — rather than 5, and is clearly above 3 because the bulk is curated specifics Claude cannot be assumed to have already pinned down. | 4 / 5 |
Actionability | Mostly executable, specific guidance ("Y: 12–24px → 0", "Scale: 0.98 → 1", "Trigger when section is ~20–30% visible", "Clamp device pixel ratio ... (1–2)", "standard offsets (8/16/24px)") plus a concrete output checklist, matching anchor 4. Not 5 because some sections stay high-level ("Morph / shared element ... Requires consistent geometry + measured layout", "Use timelines (or keyframes) with labeled beats") with no recipe, and no code recipe is provided inline for common cases; well above 3 since the common cases are covered with directly usable numbers. | 4 / 5 |
Workflow Clarity | There is a coherent progression (goals → style traits → primitives → defaults → patterns → performance → accessibility → implementation) ending in an operational two-step procedure: "What to ask the user" followed by the numbered "Output format" deliverables (1–5), with an explicit reduced-motion checkpoint ("Ask the user: 'Do you want a reduced-motion mode ...'"). Not 5 because there are no validation/feedback checkpoints and the output format is not explicitly tied to the user's answers; not 3 because no destructive or batch operations are involved, so the workflow cap does not apply, and the sequence plus checkpoints are present. | 4 / 5 |
Progressive Disclosure | A single ~185-line file with zero references: detailed catalogs ("Choreography patterns", the duration/easing/stagger "Defaults" tables, hover/scroll specs) are inlined when a skill this size would split them into reference files, matching anchor 3 (good structure, but content that should be separate is inline). Not 4 because there is no offloading or navigation to further detail at all; not 2 because sections are clearly headed and nothing is buried or monolithic. | 3 / 5 |
Total | 15 / 20 Passed |