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.
An exceptionally well-structured instruction-only skill: terse, concrete, and explicitly sequenced with validation checkpoints and real one-level-deep references. Its only shortcoming is a couple of dense run-on paragraphs that slightly tax the reader.
Suggestions
Split the dense Step 3 paragraph into a short bulleted list of true-stop conditions so the failure-mode guidance ('stopping the run there is the primary failure mode') stands out.
Break the long 'Authority comes from the babysit invocation' boundary sentence into labeled Downward/Upward sub-points for faster scanning.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense and operational with no padding or explanation of concepts Claude already knows, assuming competence throughout; a few run-on sentences (notably the Step 3 paragraph) could be tightened, keeping it just below the lean anchor 5. | 4 / 5 |
Actionability | Provides concrete executable commands ('gh run rerun <run-id> --failed -R <host>/<owner>/<repo>', 'gh stack merge', 'ce-resolve-pr-feedback mode:pipeline') and exact stop-condition predicates (MERGEABLE, CLEAN, open_needs_human == 0) covering the common cases. | 5 / 5 |
Workflow Clarity | A clearly sequenced four-step process with an explicit ordering invariant, validation checkpoints (terminal check, stale-SHA cancellation), enumerated stop conditions, and feedback loops for the destructive/batch-style push and branch mutations. | 5 / 5 |
Progressive Disclosure | SKILL.md is a concise overview with well-signaled one-level-deep references to real bundled files (references/tick.md, envelope.md, settle.md, etc.) and the pr-snapshot script; content is appropriately split and easy to navigate. | 5 / 5 |
Total | 19 / 20 Passed |