Content
100%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-crafted instruction skill body: lean, concrete, and properly progressive-disclosed. The decision table and hard rules give actionable guidance while templates and patterns are correctly offloaded to clearly labeled references.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~42-line body is lean and assumes Claude's competence — a decision table, concrete hard rules, and a signaled reference list — with no padding explaining what a PR is or how tooling works. | 3 / 3 |
Actionability | Concrete, actionable guidance for an instruction skill: a change→shape decision table and exact directives ("Do not include `## Summary`…", "Add a `## Changelog` section only for `feat:` and `fix:` PRs with user-visible changes") rather than vague direction; code absence is not penalized per the instruction-skill note. | 3 / 3 |
Workflow Clarity | Clear sequenced flow — match the change to a shape, open that shape's reference, then draft — with a default fallback ("default to the focused fix"); the Hard Rules list serves as the verification checklist, and no destructive/batch operation requires an explicit feedback loop. | 3 / 3 |
Progressive Disclosure | A concise overview points to five well-signaled, one-level-deep references (body-patterns, visual-patterns, examples, changelog-entries, github-pr-operations), each described by purpose; all referenced files exist and content is appropriately split. | 3 / 3 |
Total | 12 / 12 Passed |