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 exemplary instruction-only skill body: lean, rule-dense, with a clear decision gate, sequenced steps, explicit validation and rollback checkpoints, and a well-structured one-level-deep reference. The only notable gap is that executable commands and the strategy choice mechanics live entirely in the reference file.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~40-line body is all non-obvious operational rules ("Keep each PR reviewable in about ≤60 minutes", "Treat polluted diffs as base bugs"), with zero explanation of concepts Claude already knows and no padded sections. | 5 / 5 |
Actionability | Concrete thresholds (400 changed lines, ≤60 minutes), a condition-to-action decision table, and exact branch-targeting rules make the guidance mostly executable; minor gaps remain because the gh commands live only in the reference file and "when none is cached" is unexplained. | 4 / 5 |
Workflow Clarity | The six Execution Steps are clearly sequenced with explicit validation ("Verify each PR independently: CI/tests/docs/manual checks, rollback scope, and clean diff"), an error-recovery loop ("retarget or rebase until only the current work unit appears"), and a hold checkpoint (tracker stays draft until all children integrate), so the batch-operation validation cap does not apply. | 5 / 5 |
Progressive Disclosure | The body is a lean overview with a single well-signaled, one-level-deep reference whose contents are accurately described ("strategy diagrams, PR body section, branch commands, and reviewer guidance") and verified to exist in references/chaining-details.md; details are appropriately split out. | 5 / 5 |
Total | 19 / 20 Passed |