Content
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a strong, highly actionable skill that provides specific, concrete design guidance Claude wouldn't inherently possess — exact font names, hex values, CSS properties, and anti-pattern replacements. The workflow is clear with a sensible prioritization system and safety constraints. The main weakness is its length; the exhaustive checklists, while valuable, make it a large monolithic document that would benefit from being split into referenced sub-files for better token efficiency.
Suggestions
Split the detailed audit checklists (Typography, Color, Layout, etc.) into separate referenced files (e.g., AUDIT_TYPOGRAPHY.md, AUDIT_LAYOUT.md) and keep SKILL.md as a concise overview with the workflow, priority order, and rules.
Remove items that explain concepts Claude already knows well (semantic HTML tags, alt text purpose, inline styles being bad) and focus only on the design-specific judgment calls that add unique value.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is quite long (~300+ lines) and some sections border on exhaustive checklists that could be trimmed. However, most items are genuinely useful design-specific knowledge that Claude wouldn't inherently know (specific font names, specific hex values, specific CSS properties). Some items like explaining what semantic HTML is or what alt text does are things Claude already knows, but the majority earns its place. | 2 / 3 |
Actionability | The skill provides highly specific, concrete guidance throughout: exact CSS properties (`font-variant-numeric: tabular-nums`), specific hex values (`#0a0a0a`, `#121212`), named font alternatives (`Geist`, `Outfit`, `Cabinet Grotesk`), exact techniques (`scale(0.98)`, `translateY(1px)`), and specific anti-patterns with their replacements. Every checklist item pairs a problem with a concrete fix. | 3 / 3 |
Workflow Clarity | The skill has a clear 3-step workflow (Scan → Diagnose → Fix), a prioritized fix order (7 levels from lowest risk to highest polish), and explicit rules about not breaking functionality and testing after every change. The 'Fix Priority' section serves as a validation checkpoint by ordering changes by risk level. The constraint 'Do not break existing functionality. Test after every change' provides the necessary feedback loop. | 3 / 3 |
Progressive Disclosure | The content is well-organized with clear section headers and logical grouping, but it's a monolithic document that could benefit from splitting detailed audit checklists into separate reference files. The Typography, Color, Layout, Interactivity, Content, Component, Iconography, and Code Quality sections could each be their own referenced file, keeping SKILL.md as a lean overview with the workflow and priority order. | 2 / 3 |
Total | 10 / 12 Passed |