Content
82%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 solid, highly actionable body: concrete commands, a ready-to-use template, and precise checklist rules tied to diff categories. The main issues are a dangling reference to a nonexistent resources/implementation-playbook.md, the absence of any verification checkpoint in the workflow, and a small amount of generic boilerplate in Limitations.
Suggestions
Fix the Resources section: either add the referenced resources/implementation-playbook.md to the bundle or remove the pointer, since the file does not currently exist.
Add a lightweight verification step to the workflow, e.g. after step 2 confirm every file from the --stat output is categorized before generating the description.
Trim the generic Limitations boilerplate and the When-to-Use section that duplicates the frontmatter description to reclaim tokens.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is efficient — concrete commands, a compact template, and a dense checklist table — but the "Limitations" section is generic boilerplate ("Use this skill only when the task clearly matches the scope described above") and "When to Use" restates the frontmatter description. Not 5: these padded sections could be trimmed without losing skill-specific value. | 4 / 5 |
Actionability | Fully executable guidance throughout: `git diff <base>...HEAD --stat`, a copy-paste-ready markdown template with worked example rows, specific per-category checklist items, and concrete split commands (`git checkout -b feature/part-1`, `git cherry-pick`) with explicit thresholds (20 files / 1000 lines). Not 4: there are no gaps in the common-case coverage. | 5 / 5 |
Workflow Clarity | A clear 5-step numbered workflow with concrete commands and unambiguous outputs. Not 5: no validation or verification checkpoint exists (e.g., confirming every changed file from the --stat output is categorized before generating); not 3: the sequence is far more concrete than the bare-steps anchor-3 example, and the non-destructive nature of the task makes missing checkpoints a minor rather than disqualifying gap. | 4 / 5 |
Progressive Disclosure | Well-organized sections with content appropriately placed inline (template, checklist table) and a single, clearly signaled one-level reference in a Resources section. Not 5: the sole referenced file `resources/implementation-playbook.md` does not exist in the bundle — the pointer is dangling, so navigation to the referenced material fails. | 4 / 5 |
Total | 17 / 20 Passed |