Content
79%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-structured, concise skill that provides clear actionable guidance for auditing UI code against a design system. Its main strengths are token efficiency and a concrete output format with severity classification. The primary weakness is the lack of a re-validation step after auto-fixing violations, and the referenced design-rules.md file is not available in the bundle to verify.
Suggestions
Add a re-validation step after auto-fixing: re-run the audit on fixed files to confirm violations are resolved before reporting completion.
Include the `references/design-rules.md` file in the bundle, or note what to do if it's missing.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is lean and efficient. It doesn't explain what glassmorphism is, what DaisyUI is, or how F#/Fable works—it assumes Claude knows all of this. Every section serves a clear purpose with no padding. | 3 / 3 |
Actionability | Provides a concrete git command for file discovery, specifies exact file paths to read (DesignSystem.fs, design-rules.md), lists all 9 rule categories explicitly, and gives a complete report template with table format and severity classification. The example violation row is copy-paste ready. | 3 / 3 |
Workflow Clarity | The 6-step workflow is clearly sequenced and logical. However, there are no validation checkpoints or feedback loops—for instance, after auto-fixing violations there's no step to re-run the audit to verify fixes were correct. For an audit workflow that offers to auto-fix code, a re-validation step is important. | 2 / 3 |
Progressive Disclosure | The skill references `references/design-rules.md` for the full rule set, which is good progressive disclosure. However, no bundle files are provided, so we can't verify this reference exists. The 9 rule categories are listed inline as a summary which is appropriate, but the skill could benefit from clearer signaling of what's in the referenced file versus what's here. | 2 / 3 |
Total | 10 / 12 Passed |