Content
50%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill provides a reasonable process-oriented guide for creating ESLint plugin rules with clear sequencing and useful file references. However, it lacks executable code examples (no AST visitor patterns, no test case templates), has no validation/verification steps after implementation, and could be more concise by trimming repeated cautionary language about not making assumptions.
Suggestions
Add a concrete code example showing a minimal create() function with an AST visitor method and a corresponding test case, so Claude has a copy-paste-ready template rather than needing to read referenced files.
Add a validation step at the end of the workflow: e.g., 'Run `npm test -- --filter=<rule-name>` to verify all test cases pass. If tests fail, review the rule logic and fix before proceeding.'
Trim the repeated cautionary language in steps 2-3 ('Do not make assumptions', 'ask for clarification') into a single concise note at the top, freeing tokens for more actionable content.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is mostly efficient but includes some unnecessary verbosity, particularly in steps 2-3 where it repeatedly emphasizes asking for clarification and not making assumptions. Some sentences could be tightened (e.g., 'The script should be used every time a new rule is created' is redundant given the process already says to run it). | 2 / 3 |
Actionability | The skill provides concrete file paths and references to existing rules as examples, which is helpful. However, it lacks executable code snippets—there's no example of what a create() function looks like, no example AST visitor pattern, and no example test case structure. The guidance is specific in terms of process but incomplete in terms of implementation details. | 2 / 3 |
Workflow Clarity | The 5-step process is clearly sequenced and logically ordered, with explicit gates (e.g., 'Do not proceed until...'). However, there are no validation checkpoints—no step to verify the rule works, no linting test command, no feedback loop for fixing issues after writing the rule. For a multi-step code generation workflow, the absence of a 'run tests and verify' step is a notable gap. | 2 / 3 |
Progressive Disclosure | The skill references specific files (custom-rule.ts, no-v7-imports.ts, scaffold-new-rule.mjs) which is good for navigation, but no bundle files are provided to support these references. The content is somewhat monolithic—the detailed process steps, directory placement, and implementation guidance are all inline rather than structured with clear sections pointing to supporting documentation. | 2 / 3 |
Total | 8 / 12 Passed |