Content
90%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.
A lean, highly actionable instruction skill with concrete commands, real API references, and a clear workflow backed by a review checklist. Validation is present but the fix-retry feedback loop could be more explicit, and cross-skill loads are not centralized.
Suggestions
Make the snapshot/test failure feedback loop explicit: e.g., 'If a snapshot mismatches, update the fixture or rule and re-run `just test-lintrule` until it passes.'
Centralize the cross-skill load directives (diagnostics-development, testing-codegen, syntax-text-handling, eslint-migrate-options, changeset) into a single 'Related skills' section so navigation is one glance.
Consider moving the detailed Options integration-point checklist and the Actions test list into a referenced file to keep the overview shorter and improve progressive disclosure.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense and lean with no basic-concept padding; it assumes Claude's competence with Biome/Rust and points to the canonical guide ('Read only the sections relevant to the rule rather than loading the entire guide') instead of duplicating it. | 5 / 5 |
Actionability | Provides copy-paste-ready commands ('just new-js-lintrule useMyRule', 'just test-lintrule', 'just gen-rules', 'just f', 'just l') alongside concrete API names (Ast<Node>, Semantic<Node>, run(), action(), FixKind::Safe, declare_lint_rule!) and exact file paths. | 5 / 5 |
Workflow Clarity | A clear 6-step sequence plus an explicit Review Checklist and snapshot inspection as validation checkpoints, but the error-recovery feedback loop (snapshot fails -> fix -> re-run) is only lightly implied rather than spelled out. | 4 / 5 |
Progressive Disclosure | Well-organized into clear sections with one-level-deep references to canonical docs and other skills, all clearly signaled; minor gaps because cross-skill loads are scattered through the body and some detail (Options integration points, Actions test list) is inlined rather than split out. | 4 / 5 |
Total | 18 / 20 Passed |