Content
86%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 tight, well-structured, mostly executable skill that assumes Claude's competence and adds only Flow-specific knowledge. The main gap is the absence of a fix-and-recheck feedback loop and an unspecified error-reporting format.
Suggestions
Add a closing feedback loop: after reporting errors, fix them and re-run `yarn flow $ARGUMENTS` until clean.
Specify the error-reporting format in step 2, e.g. group by file with line numbers and the Flow error code.
Clarify when to prefer `yarn flow-ci` over `yarn flow` beyond 'comprehensive (slow)' — e.g., pre-merge checks vs local iteration.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and efficient throughout — no padding or explanation of concepts Claude already knows; the 'Common Mistakes' section earns its place with Flow-specific guidance ($FlowFixMe suppressions, type imports). | 5 / 5 |
Actionability | Provides copy-paste-ready commands (`yarn flow $ARGUMENTS`, `yarn flow-ci`) and a concrete renderers table, but step 2 ('Report type errors with file locations') gives no output format or example, leaving a minor gap. | 4 / 5 |
Workflow Clarity | A clear numbered sequence (run, report, optional flow-ci) for a simple read-only check, but there is no closing feedback loop (fix flagged errors then re-run) to confirm resolution. | 4 / 5 |
Progressive Disclosure | Under 50 lines, no external references needed, and well-organized into clearly labeled sections (Renderers, Instructions, Common Mistakes), satisfying the simple-skill exception for progressive disclosure. | 5 / 5 |
Total | 18 / 20 Passed |