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 excellent content body: fully executable commands, a well-sequenced workflow with explicit validation and approval gates, and tight organization with no bundle-file overhead. The only weakness is minor redundancy between the intro, the frontmatter description, and the red-flags section.
Suggestions
Trim the opening paragraph, which restates the frontmatter description and duplicates two of the 'Red flags / don't' rules; the sections below already carry that content.
The 'When to use' section is the natural place for the trigger guidance missing from the description; cross-checking them would keep both consistent.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and command-first with essentially no explanation of concepts Claude already knows; however, the opening paragraph repeats the frontmatter description nearly verbatim and pre-states two rules ('never act on title or labels alone', approval gating) that the 'Red flags / don't' section restates. These minor trimmable redundancies fit anchor 4 rather than 5's 'every token earns its place'. | 4 / 5 |
Actionability | Fully executable, copy-paste-ready commands throughout: exact 'gh pr list/view' invocations with --json field lists, a 'git merge-tree' conflict check, trailer formats with a gh api command for the noreply id, and a 'cargo fmt && cargo test' sanity check. Placeholders like <release-branch> are appropriately repo-specific, so the common cases are covered as anchor 5 requires. | 5 / 5 |
Workflow Clarity | A clear 8-step sequence with explicit validation checkpoints: read-everything-decide-nothing pull, code+tests+comments+checks confirmation, merge-tree test against the real landing branch, trailer validation, and a local fmt/test sanity check before recommending. Risky outward actions (merge/close/tag/post) are explicitly gated on maintainer approval, and the merge-tree step includes an error-recovery rationale. This matches anchor 5, not 4, because validation is explicit and repeated rather than implied. | 5 / 5 |
Progressive Disclosure | No bundle files exist (no references/, scripts/, or assets/), and none are needed: the skill is a single well-sectioned file (~100 lines) with no nested references and no inline content that clearly belongs in a separate file. References to .github/AUTHOR_MAP and auto-close-harvested.yml are repo artifacts, not skill bundle references. Per the simple-skill scoring note, well-organized sections with no need for external references score 5. | 5 / 5 |
Total | 19 / 20 Passed |