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.
A well-engineered skill body: fully executable commands, a four-phase workflow with explicit validation and an error-recovery feedback loop, and exemplary progressive disclosure via a routing table to five real, one-level-deep reference files. The only weakness is mild redundancy between the when-to-use section, the safety rules, and the workflow steps.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude's competence — no padding about what PRs or CI are — with every section operational (commands, routing table, safety rules). Minor trimmable redundancy: "When to use this skill" restates the frontmatter description, and the Git safety rules repeat steps already in Phase 3 ("Always stash", "Always verify"), so it falls just short of the every-token-earns-its-place level-5 anchor. | 4 / 5 |
Actionability | Copy-paste-ready commands throughout: `gh pr list --state open --json number,title,headRefName,statusCheckRollup,reviewDecision,mergeable --limit 30`, `gh pr checks <number>`, `git checkout -B <branch> github/<branch>`, plus a concrete commit format `fix(<scope>): 🔧 <english description>` with real scope examples. The only uncommanded steps ("Reproduce the issue locally (build, test, or lint)") are project-specific and appropriately delegated to references, matching the level-5 anchor of executable commands covering common cases. | 5 / 5 |
Workflow Clarity | Four clearly sequenced phases with explicit validation ("Reproduce the issue locally", "Verify locally: build → test → lint", post-push `gh pr checks`), a genuine feedback loop ("If CI still fails, loop back to Phase 3 for that PR"), and a closing checklist — the exact shape of the level-5 anchor. The batch-operation validation cap does not apply because local verification and CI re-checking are both present; it is above level 4 because the error-recovery loop and self-check checklist are explicit rather than implied. | 5 / 5 |
Progressive Disclosure | The SKILL.md is a concise overview with a dedicated routing table mapping each task to its reference file; all five referenced files (discovery, triage, fix-workflow, ci-pipeline, fix-recipes) exist in `references/`, and inspection confirms they are one level deep (a single sibling cross-reference, no nested chains). This matches the level-5 anchor of a clear overview with well-signaled, one-level-deep references and easy navigation. | 5 / 5 |
Total | 19 / 20 Passed |