Content
87%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is concise, highly actionable, and well-organized with concrete paths and token patterns. Its main weakness is the absence of a validation/verification checkpoint in the fix workflow, which is important since fix=true applies destructive edits.
Suggestions
Add a validation step after applying fixes, e.g. 'After applying fixes: run typecheck/build and confirm components render correctly before finishing.'
Clarify the analysis step with a brief checklist tying each anti-pattern back to the corresponding rule section.
State explicitly what to do when the barrel file or referenced source files are missing or out of date.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and token-efficient: dense tables, token lists, and terse rules with no explanation of concepts Claude already knows and no padding. | 3 / 3 |
Actionability | Guidance is concrete and copy-paste ready — exact file paths (e.g. 'apps/sim/components/emcn/index.ts'), exact token patterns ('text-[var(--text-primary)]'), and a full intent-to-variant table; for an instruction-only review skill this is highly actionable. | 3 / 3 |
Workflow Clarity | A clear numbered sequence exists (read barrel, read globals.css, analyze, apply/propose fixes), but applying fixes is a potentially destructive operation with no validation or verification checkpoint (e.g. typecheck, build, or visual check), which caps workflow clarity at 2. | 2 / 3 |
Progressive Disclosure | Content is organized into well-signaled one-level-deep sections with no nested references; the few external pointers (e.g. '.claude/rules/emcn-components.md') are single-level, and the single-file structure is appropriate for a focused review skill. | 3 / 3 |
Total | 11 / 12 Passed |