Content
96%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 an exemplary operational skill: fully executable commands, a tightly sequenced loop with validation checkpoints and recovery paths, and lean token usage throughout. The only structural improvement would be splitting edge-case sections into reference files to shorten the main SKILL.md.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes competence throughout — no concept explanations, only repo-specific procedure — and even the incident citations ("this left two rounds unanswered on PR #1097", "#8288 patched one resolver about 25 times") are single-clause justifications for rules, matching anchor 5 rather than the trimmable over-explanation of anchor 4. | 5 / 5 |
Actionability | Guidance is fully executable: exact gh commands with explicit --json fields, complete jq pipelines (including the "Bind `.id as $id` first" gotcha), a copy-paste changeset template, and a guarded merge command with concrete path exclusions, matching anchor 5. | 5 / 5 |
Workflow Clarity | The "Each tick" loop is a clear numbered sequence with explicit validation checkpoints (rerun the unreplied-comments query before merging, confirm HEAD equals headRefOid) and error-recovery feedback loops (retry on failed queries, merge-and-repush on rejection), matching anchor 5; the destructive-operation cap does not apply since validation is pervasive. | 5 / 5 |
Progressive Disclosure | Sections are well-organized and cross-skill delegation is clearly signaled ("Ownership … all live in `ship`"; "see `factory-run/references/babysit-round.md`"), but everything is inline in a ~160-line file — edge-case procedures like the changeset and pnpm-lock-conflict handling could be split into reference files — so it fits anchor 4 rather than the appropriately split content of anchor 5. | 4 / 5 |
Total | 19 / 20 Passed |