Content
92%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 is a tightly structured, validation-rich triage workflow with concrete commands, paths, and label values; it assumes competence and avoids padding. The only minor gap is the absence of a worked example showing a complete verdict and label set.
Suggestions
Add a short worked example showing a representative final output (e.g. a one-paragraph verdict plus the resulting Type/Priority/Difficulty labels) to lift actionability to fully copy-paste-ready.
Optionally show the exact gh CLI incantation for fetching issue body/comments/labels so the 'Read current evidence' step is directly executable rather than descriptive.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~50-line body is lean, assumes Claude's competence (no explanations of what an issue or GitHub is), and every line is actionable guidance or a checkpoint with no padding. | 5 / 5 |
Actionability | Concrete commands and paths are given throughout (`--help`, `docs/openapi/latest.json`, the `sosumi.ai` mirror, `CONTRIBUTING.md`, `$develop-asc-change`, exact label values, RED-GREEN test plan), but no worked example of a final verdict/output shape is provided. | 4 / 5 |
Workflow Clarity | A clearly sequenced process (Read evidence → Classify → Apply labels → Define readiness → Automation contract) with explicit checkpoints (reproduce, check already-fixed/duplicate/unsupported, remove conflicting labels, don't close until merged+verified) and checklists (label buckets, verdict types, readiness list). | 5 / 5 |
Progressive Disclosure | Under 50 lines with no bundle files to reference, organized into five well-signed section headers; the inline pointers (docs/openapi/latest.json, CONTRIBUTING.md) are repo artifacts appropriately kept inline rather than skill-bundle references. | 5 / 5 |
Total | 19 / 20 Passed |