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 strong, execution-ready instruction set: concrete gh commands, a complete output template, a comprehensive classification/routing table, and a tightly sequenced workflow with stop conditions and a final checklist. Its only flaws are minor redundancy in the opening scope statements; progressive disclosure is clean since the skill needs no external files and none are dangling.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and assumes Claude's competence — exact gh commands with JSON field lists, a fill-in comment template, and a decision table with no padding or explanations of known concepts. It is not a 5 because the opening two paragraphs restate the same scope ('classify the issue, choose the next route, post or update one concise issue comment... and stop' vs 'classify the issue, explain the route, update the issue, and stop') and 'Apply only the labels listed above' repeats the Label policy intro. | 4 / 5 |
Actionability | Guidance is copy-paste ready throughout: exact commands ('gh issue view "$ISSUE" --comments --json ...', a parameterized GraphQL query, 'gh issue list --search ...'), an exact comment shape with enumerations for every field, precise label names, and a type/route/approval table covering all common cases. Not below 5 — nothing is pseudocode or abstract. | 5 / 5 |
Workflow Clarity | Five clearly numbered steps with explicit stop conditions ('if the input cannot be resolved, state what is missing and stop', 'if the issue is not open, stop'), validation checkpoints in the gather phase (issue state, existing discussion, closing PRs, duplicates), duplicate-comment handling, and a literal checklist in step 5. The skill processes a single issue and explicitly forbids closing it, so it is neither destructive nor batch and the validation cap does not apply. | 5 / 5 |
Progressive Disclosure | No bundle files (references/scripts/assets) exist and none are referenced — the only cross-references are to sibling skills, not files — so there are no orphaned or nested references. The single-file instruction set is appropriately sized with clear, well-signaled sections (Label policy, Output contract, Workflow) and easy navigation; nothing that belongs in a separate file is inlined. | 5 / 5 |
Total | 19 / 20 Passed |