Content
92%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 strong, highly actionable skill body: a gated five-step workflow with measurable SSIM validation and an error-recovery loop, copy-paste commands, a concrete output contract, and exemplary progressive disclosure via a pattern-to-reference table with verified bundle files. The only weakness is redundancy — the one-way/Remotion-only boundary is repeated across three sections and the @remotion/lambda clarification appears twice.
Suggestions
Consolidate the one-way/Remotion-only boundary statement into a single section — the opening blockquote, the 'NOT SUPPORTED' subsection, and 'What this skill explicitly does NOT do' all repeat the same reverse-export/non-Remotion exclusions; keeping it once (plus the frontmatter description, which already covers it) would trim ~10 lines.
Cut the parenthetical @remotion/lambda re-clarification at the end of 'What this skill explicitly does NOT do' — it duplicates the Step 1 warning-case explanation and the escape-hatch.md link already given.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is efficient — it assumes Claude knows React/GSAP/Remotion basics, explains nothing elementary, and keeps every section operational. However, the one-way/Remotion-only boundary is stated three times (the opening blockquote, the 'NOT SUPPORTED' section under 'When to use', and 'What this skill explicitly does NOT do'), and the @remotion/lambda-is-a-warning clarification appears both in Step 1 and again at line 108. This matches the level-4 anchor (efficient with minor instances that could be trimmed) rather than level-5 (every token earns its place), since consolidating the boundary statement to one section would remove real duplication. | 4 / 5 |
Actionability | Guidance is fully executable: a copy-paste render/diff command block ('npx remotion render <CompositionId> out/baseline.mp4', 'npx hyperframes render --skill=remotion-to-hyperframes', '../../scripts/render_diff.sh'), a concrete output spec (root div with 'data-composition-id, data-start="0", data-duration... data-fps, data-width, data-height', 'one paused gsap.timeline({paused: true})', 'window.__timelines["<composition-id>"] = tl;'), and exact validation prerequisites ('Config.setVideoImageFormat("png")' + 'Config.setColorSpace("bt709")'). Matches the level-5 anchor — copy-paste ready commands covering the common cases. | 5 / 5 |
Workflow Clarity | The five-step workflow has an explicit gate (Step 1 lint: 'If any blocker fires, stop'), an explicit validation step with a measurable threshold ('~0.02 below p05 of the source's complexity tier'), and a genuine feedback loop for error recovery ('If the diff fails, run scripts/frame_strip.sh to see which frames diverged, then re-read the relevant timing/sequencing/media reference'). Matches the level-5 anchor: clear sequence, explicit validation, and retry guidance after failure. | 5 / 5 |
Progressive Disclosure | The SKILL.md body is a lean overview with well-signaled, one-level-deep references: a table maps each source pattern ('Sequence, Series, Loop...' → sequencing.md; '@remotion/lottie' → lottie.md) with the instruction 'Don't load all of them — load only what the specific source needs', api-map.md serves as an in-scope index, and all ten referenced files plus scripts/ and assets/test-corpus/ exist in the bundle as described. Matches the level-5 anchor — clear overview, appropriately split content, easy navigation. | 5 / 5 |
Total | 19 / 20 Passed |