Content
71%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 delivers a concrete, well-sequenced workflow with specific file paths and sensible gating on user input. Its main weaknesses are mild redundancy in the step text and a missing verification step after tests are written.
Suggestions
Add a final verification step — e.g. '6. Run the rule's tests (e.g. via the package's test command) and fix the rule or tests until they pass' — to close the workflow with a feedback loop.
Tighten step 1 to remove the duplicated justification ('The script should be used every time... will ensure the same format') into a single clause, and merge the repeated ask-for-clarification guidance in steps 2 and 3.
Show how to invoke the scaffold script (command line) and a minimal valid/invalid test case example so the guidance is copy-paste ready.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient, project-specific guidance, but contains avoidable padding: 'The script should be used every time a new rule is created. This script will ensure that the same format is used for all custom rules created' says the same thing twice, and the ask-for-clarification instruction is repeated across steps 2 and 3. This fits 'mostly efficient but could be tightened' rather than 'minor instances of over-explanation'. | 3 / 5 |
Actionability | Guidance is concrete and executable for an instruction-only skill: exact script path (packages/eslint-plugin-cds/scripts/scaffold-new-rule.mjs), reference files to consult (custom-rule.ts, no-v7-imports.ts), and explicit directory placement. Minor gaps — the script invocation syntax is not shown and no snippet of a create() visitor or test case is included — keep it below 'fully copy-paste ready'. | 4 / 5 |
Workflow Clarity | A clear numbered 5-step sequence with input-gating checkpoints ('Do not proceed until the user has provided at least one valid and one invalid example') is present. However, there is no final validation step — the tests written in step 5 are never run against the rule — a minor validation gap that fits 'clear sequence with most checkpoints present'. | 4 / 5 |
Progressive Disclosure | The skill is under 50 lines with no need for external references (no references/, scripts/, or assets/ bundle files exist; the paths cited are repo files), and it is organized into two well-signaled sections ('Process', 'Directory placement'). Per the rubric's simple-skill exception, this earns the top progressive disclosure score. | 5 / 5 |
Total | 16 / 20 Passed |