Content
67%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 well-engineered routing-and-contract skill body: concrete numeric budgets, a phased workflow with per-phase reads, and exemplary reference navigation. Its weaknesses are redundancy — version matrices and licensing notes repeated across the comment block, Trigger Guidance, Core Contract, and routing rules — and inlined detail that belongs in the reference files it already points to.
Suggestions
Remove the CAPABILITIES_SUMMARY HTML comment (or the overlapping Trigger Guidance section) — they duplicate the same browser-support and capability data, and the version matrices differ slightly between the two copies, which risks drift.
State the GSAP licensing note (all plugins free since the Webflow acquisition) exactly once, in the library-selection routing row, instead of three times across the comment block, Core Contract, and routing rules.
Move the full browser version matrices (View Transitions, @starting-style, interpolate-size, animation-trigger) into reference/modern-css-animations.md and keep only the feature names plus a single 'see reference for support matrix' pointer in the body.
Consolidate the scroll-triggered vs scroll-driven distinction, which is currently explained in three places (Trigger Guidance, Core Contract, and the scroll recipe note), into one canonical definition.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with non-obvious specifics (frame budgets, browser support matrices, CLS/INP thresholds), but the CAPABILITIES_SUMMARY HTML comment nearly duplicates the Trigger Guidance, the View Transitions and scroll-driven version matrices appear twice with slightly different data, the GSAP licensing note appears three times, and the scroll-triggered vs scroll-driven distinction is explained twice — more than the 'minor instances' of the anchor 4, so it fits 'mostly efficient but could be tightened'. | 3 / 5 |
Actionability | Guidance is highly concrete for an instruction/routing skill — exact thresholds (150-300ms transitions, ≤2 will-change elements, ≤4 parallax layers at ≤120px, animation-duration: 1ms for Firefox), named APIs (interpolate-size, animation-trigger, @starting-style), a usable rAF→setTimeout pattern, and exact journaling formats. It stops short of anchor 5 because no executable code examples are inline (all deferred to reference files) and some patterns (e.g., linear() spring approximation) are named without any illustrative curve. | 4 / 5 |
Workflow Clarity | The SURVEY → PLAN → VERIFY → PRESENT table explicitly maps each phase to a required action, key rule, and file to read, and the VERIFY phase names concrete checks (reduced-motion and performance validation). It misses anchor 5 because there is no error-recovery feedback loop (what to do when verification fails) — checkpoints exist but retry/fix guidance is implicit. | 4 / 5 |
Progressive Disclosure | Navigation is strong: a Reference Map with explicit "Read this when" conditions, Read-First columns in the Recipes table, and Read-next routing keep references one level deep and clearly signaled. It falls short of anchor 5 because the ~250-line body inlines substantial detail — full browser version matrices and per-recipe constraints — that thematically belongs in reference/modern-css-animations.md and the individual recipe references; no bundle files were provided to verify, so this is scored on the reference structure alone. | 4 / 5 |
Total | 15 / 20 Passed |