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.
The body is a well-structured, highly actionable design doctrine with an unusually strong workflow: an explicit sequence, tiered verification, and a genuine diagnose→repair→re-score feedback loop. Its main weakness is that everything lives inline in one large SKILL.md with no reference files, costing token budget on every invocation, plus minor redundancy between overlapping sections.
Suggestions
Move the low-frequency detail blocks into one-level-deep reference files (e.g. references/slop-diagnostic.md, references/deck-rules.md, references/format-standards.md) and keep short summaries plus explicit links in SKILL.md, so the ~650-line body shrinks to an overview.
Deduplicate overlapping guidance: the "When To Use" list restates the top decision table, and the "Portable Opening Prompt Pattern" paragraph re-summarizes the entire skill — both can be trimmed to a pointer or removed.
Make abstract composition guidance concrete with one- or two-line inline examples (e.g. what 'rhythm via scale and interruption' looks like in a section), mirroring how the Surface-First section already pairs rules with concrete instances.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is directive and assumes competence (no explanation of what CSS or HTML is; every section issues instructions rather than tutorials), with only minor trimming opportunities — the "When To Use" list overlaps the top decision table, and the "Portable Opening Prompt Pattern" re-summarizes the whole skill. This fits 'efficient; minor instances of over-explanation that could be trimmed' rather than the fully lean anchor 5. | 4 / 5 |
Actionability | Concrete thresholds and templates abound — "Default slide size: 1920×1080, 16:9", "Mobile hit targets should be at least 44px", "For print documents, text should be at least 12pt", exact versioning filenames ("Name v2.html"), a copy-paste final-response example, and a ten-tell diagnostic with per-tell repair mapping. Not anchor 5 because some guidance stays abstract (e.g. "Design with rhythm: scale, whitespace, density, alignment...") without examples of what rhythm looks like in practice. | 4 / 5 |
Workflow Clarity | An explicit 8-step workflow (understand brief → gather context → commit to surface → define system → choose format → build → verify → report) includes an explicit Verify step with tiered minimums, and the Slop Diagnostic adds a true feedback loop: score out of 10, repair matched to the diagnosis, then "Re-score after repairing. Do not declare done while compositional tells (3, 8, 10) are still firing". This matches the anchor requiring explicit validation steps and feedback loops. | 5 / 5 |
Progressive Disclosure | The skill has no bundle files at all: all ~650 lines are inline in SKILL.md under well-organized section headers. Long blocks that could live in one-level-deep reference files (the ten-tell Slop Diagnostic, Deck/Prototype rules, Typography/Color doctrine, hosted-tool remap list) are inlined, fitting 'content that should be separate is inline' — anchor 3, not 4, because there are no references to be clearly signaled and the overview-to-detail split the rubric expects is absent. | 3 / 5 |
Total | 16 / 20 Passed |