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.
Excellent workflow skill: every step is concrete and executable, validation checkpoints and bounded feedback loops are explicit at each stage, and the body stays lean with no padding or over-explanation. The only trims are a single justificatory sentence and slightly sparse trigger synonyms in the description rather than the body.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean, imperative, and assumes Claude's competence throughout — no concept explanations, no library comparisons, no filler. One minor instance of trimmable rationale: "fixing without first reproducing the bug tends to produce incomplete fixes and regressions" justifies an instruction Claude would follow anyway. Matches anchor 4 (efficient, minor over-explanation) rather than 5 because of that sentence. | 4 / 5 |
Actionability | Fully executable guidance: a copy-paste-ready `gh issue view` command with a complete markdown template, concrete build/verify commands (`npm run build && npm run bundle`, `node dist/cli.js`), and specific artifact paths (`.qwen/issues/issue-<number>.md`). Directive steps (fix, tests, review triage) are concrete and specific, matching anchor 5 for an orchestration-style skill. | 5 / 5 |
Workflow Clarity | Six clearly sequenced steps with explicit validation gates: reproduction gate ("If the status is `NOT_REPRODUCED`, report that and stop"), verification feedback loop ("If the verification status is `STILL_BROKEN`... go back to Step 3"), a hard gate ("Do not proceed until verification returns `VERIFIED_FIXED`"), and bounded iteration rules ("Do not loop more than 3 times... without asking the user"). This matches anchor 5: explicit validation steps, error-recovery feedback loops, and loop bounds. | 5 / 5 |
Progressive Disclosure | The skill is a single self-contained file with no bundle directory, and nothing in it belongs in a separate reference file — the issue template is short and used inline. Sections are clearly headed and easy to navigate, and external pointers (`structured-debugging` skill, `test-engineer` agent, AGENTS.md self-audit) are clearly signaled one-level references, matching anchor 5's structure. | 5 / 5 |
Total | 19 / 20 Passed |