Content
77%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.
An expertly structured skill: novel domain knowledge throughout, a fully sequenced workflow with explicit validation and feedback loops, and exemplary progressive disclosure with conditional load triggers and verified one-level-deep references. The weaknesses are duplicated guidance between the workflow steps and Critical Rules (and a triple-explained fallbackNodeId), plus a couple of dangling example pointers that cost both conciseness and actionability.
Suggestions
Deduplicate guidance repeated across sections: fallbackNodeId is explained in Step 2 and twice in Step 3, and Step 5's two bullets restate Critical Rules 2 and 4 — consolidate each into one location with a cross-reference.
Replace 'See the interleaved-transform example' with an explicit link to references/examples-and-anti-examples.md (and its specific section) so the pointer is navigable.
Trim the instruction to load gotchas.md from three places (Step 3 bullet, Step 5, References index) to the single most decision-relevant point.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with genuinely novel Figma-MCP signal and assumes Claude's competence, but contains real duplication: fallbackNodeId is explained in Step 2 and then twice more in Step 3, 'load gotchas.md' is instructed three times, and Step 5's bullets restate Critical Rules 2 and 4 — fitting the 'mostly efficient but could be tightened' anchor rather than the minor-trims anchor. | 3 / 5 |
Actionability | Highly concrete — exact tool-call signatures, timing formulas (startMs + orderPosition * staggerStepMs), worked offset math ([45, 125, 125] -> [0, 80, 80]), and copy-paste transformTemplate patterns — but full worked examples are delegated to references and 'See the interleaved-transform example' is a dangling pointer without a link, leaving minor gaps. | 4 / 5 |
Workflow Clarity | Five explicitly sequenced steps (confirm static context, fetch motion data, merge, apply, validate) with a dedicated validation step, an explicit feedback loop ('reload, observe, iterate'), and pre-batching validation for what is a batch operation — matching the top anchor's clear sequence, explicit validation, and error-recovery guidance. | 5 / 5 |
Progressive Disclosure | Clear overview with six one-level-deep references that all exist on disk, each signaled with explicit 'Load when...' conditions and inline load instructions at decision points; cross-links between references are lateral siblings with back-navigation (no nested chains), and the deep-link anchor (#swiftui-translation) resolves — matching the top anchor. | 5 / 5 |
Total | 17 / 20 Passed |