Content
75%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-architected index-style SKILL.md: the routing table is token-efficient and highly navigable, the default-vs-blueprint decision rule is crisp, and the script usage is concrete with an audit step after authoring. Main gaps are a redundant restatement of the core contract, no inline example of the default composition workflow, and a routing table pointing at many files absent from the provided bundle.
Suggestions
Trim the parenthetical in Critical Constraints that re-lists hyperframes-core's rules (no Math.random / Date.now, no repeat: -1, etc.) down to a short pointer like 'Prerequisite: hyperframes-core → Non-Negotiable Rules' plus only the two animation-craft additions, since the body itself says 'Don't restate the full contract here'.
Add a 5-10 line minimal example of the default workflow (picking two rules from rules-index.md and gluing them on a single paused GSAP timeline) so the most common path is executable without loading another file.
Close the feedback loop for the audit step: state what to do when animation-map.json reports dead zones, stagger inconsistency, or lifecycle warnings (fix the timeline, re-run, only ship when clean) to raise workflow clarity.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and format-driven — a routing table instead of prose, terse decision guidance ("Pick 2-4 rules from rules-index.md, glue them together with a single paused GSAP timeline, done"), and no explanation of concepts Claude already knows. It misses level 5 because the Critical Constraints section restates a substantial chunk of hyperframes-core's contract ("single paused timeline, data-duration governs length, no Math.random / Date.now / performance.now, no repeat: -1, no page-load gsap.set...") immediately after saying "Don't restate the full contract here" — that parenthetical could be trimmed to a pointer. | 4 / 5 |
Actionability | Mostly executable guidance: a concrete script invocation ("node skills/hyperframes-animation/scripts/animation-map.mjs <composition-dir> --out <composition-dir>/.hyperframes/anim-map") with its output contract, and a routing table mapping each need to a specific file. It falls short of level 5 because the default path ("compose atomic rules") never shows even a minimal timeline-composition snippet, so the primary workflow relies on referenced files rather than copy-paste-ready guidance here. | 4 / 5 |
Workflow Clarity | The decision flow is clear and ordered: default is composing rules, explicit conditions for when to load a blueprint ("The scene matches an existing pre-designed multi-phase template... or You want runnable ground-truth code for a complex 4-5 phase choreography"), runtime selection guidance, and a post-authoring validation step ("Use it to audit choreography (dead zones, stagger consistency, lifecycle warnings) after authoring") via the animation-map script. It is not level 5 because there is no explicit feedback loop (audit → fix → re-audit) and no validation step before the audit for the authoring workflow itself. | 4 / 5 |
Progressive Disclosure | Excellent structure: a routing table covering every need with well-signaled, one-level-deep references, explicit non-speculative reading guidance ("Do not read it speculatively"), and pointers to sibling skills in See Also. It falls short of level 5 because the routing table references many paths (rules-index.md, rules/<name>.md, blueprints-index.md, blueprints/<id>.md, transitions/overview.md, transitions/catalog.md, techniques.md, adapters/*.md) that are not present in the shipped bundle — only references/motion-blur.md and scripts/ exist — so navigation is broken for most routes as bundled. | 4 / 5 |
Total | 16 / 20 Passed |