Content
78%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 lean, highly actionable instruction skill with concrete numeric guidance (space proportions, word counts, font sizes, figure limits) and a clear five-step workflow. Its main defects are the orphaned bundle files — scripts/main.py and references/guidelines.md are never referenced and partly contradict the body — and the absence of any final verification checkpoint on the delivered layout plan.
Suggestions
Reference the bundle files from SKILL.md or remove them: either link references/guidelines.md under a "## Further guidance" heading and document when to run scripts/main.py, or delete both — currently they are unreachable and main.py's 2-column output contradicts the body's 3–4 column layout.
Add a final validation step to the Core Workflow, e.g., "Step 6 — Verify the plan: check that space allocations sum to ~100%, word counts meet the Step 3 targets, and every included figure was mentioned in the user's material."
Remove the redundant blockquote after the Out-of-scope list (it repeats the point about not creating the visual design file) and the generic white-space design tip to tighten conciseness.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is efficient — domain knowledge is delivered as compact tables (space allocation percentages, per-section word targets) rather than prose, and it never pads with basics like what a poster is. Minor trims remain: the blockquote "The actual visual design file should be created in a poster design tool" duplicates the out-of-scope bullet, and "White space is not wasted — crowded posters lose viewers" is design advice Claude already holds, keeping it below the score-5 'every token earns its place' anchor. | 4 / 5 |
Actionability | As an instruction-only skill the guidance is fully actionable: a concrete section hierarchy tree, "2–4 key figures maximum", panel titles "≤8 words", per-section word-count targets, font sizes "title ≥ 72pt, headers ≥ 36pt, body ≥ 24pt", and an explicit deliverables list in Step 4 (section plan, figure plan, narrative flow note, take-home sentence). Per the rubric's scoring notes, absence of code is not penalized for instruction-only skills when the guidance is this specific and covers the common cases. | 5 / 5 |
Workflow Clarity | The Core Workflow is a clear 5-step sequence (story → figures → text compression → layout plan → design tips) with an input checkpoint ("If the source material is too sparse to create a layout, ask for the key finding and methods before proceeding"). It stops short of the score-5 anchor because there is no verification of the output — e.g., checking that space allocations sum to ~100% or that word-count targets were met before delivering — which is a minor validation gap, not the missing-validation cap (no destructive or batch operations are involved). | 4 / 5 |
Progressive Disclosure | The body itself is well-organized with clear sections, but the actual bundle is not reflected in it: references/guidelines.md (a 3-line stub) and scripts/main.py (a 2-column layout planner that contradicts the body's 3–4 column recommendation) exist in the bundle yet are never referenced, linked, or explained in SKILL.md. Per the rubric's guidance to score against the real bundle structure, this is 'references present but not clearly signaled' — the score-3 anchor rather than the score-4 'most content appropriately placed' level. | 3 / 5 |
Total | 16 / 20 Passed |