Content
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is a well-sequenced, actionable verification workflow with executable commands and clear degradation handling. Its main weakness is progressive disclosure: everything sits in one inline file with no references to split detail into.
Suggestions
Move the full report template and detailed per-dimension heuristics into a reference file (e.g. references/verify-report.md) and link to it from a concise overview in SKILL.md to improve progressive disclosure.
Tighten obvious-explanation padding such as the checkbox legend and repeated 'Add ... issue' phrasing to improve token efficiency.
Consider adding a short validation feedback loop (run verify -> review CRITICAL issues -> re-run) to make the workflow's error-recovery step explicit rather than implied.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body assumes Claude's competence and avoids explaining basic concepts, but minor padding (e.g. spelling out '- [ ] (incomplete) vs - [x] (complete)') and some restated guidance could be tightened to earn the top anchor. | 2 / 3 |
Actionability | It provides concrete, copy-paste-ready commands ('openspec list --json', 'openspec status --change ... --json', 'openspec instructions apply ...') plus a fully specified report template, matching the executable-guidance anchor. | 3 / 3 |
Workflow Clarity | The eight-step sequence is explicit and ordered, with a severity-tiered issue checklist, graceful-degradation branches, and false-positive guidance that function as validation checkpoints for a read-only verification task. | 3 / 3 |
Progressive Disclosure | No bundle files exist and the ~170-line body is a single monolith with no one-level-deep references; it is well-sectioned but content that could live in separate files (report template, heuristics) is inline, so it does not reach the top anchor. | 2 / 3 |
Total | 10 / 12 Passed |