Content
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
Highly actionable content with executable scripts, a well-sequenced destructive-batch workflow and validation gates, and clean single-file organization. The only weakness is mild verbosity in a few anecdotal asides that could be tightened.
Suggestions
Trim the recurring 'mirrors gh-bulk-repo-edit' comparisons and the personal '76 PRs' anecdote in the Pitfalls section to recover tokens without losing the substantive warning about search substring-matching.
The two paragraphs in 'The decision tree' explaining the unmergeable bucket's social trade-offs could be condensed to the core instruction plus the two legitimate options.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly lean and free of concepts Claude already knows, but contains some chatty padding — the recurring asides comparing to gh-bulk-repo-edit and the personal anecdote about the '76 PRs' count could be trimmed. | 2 / 3 |
Actionability | Provides copy-paste-ready, executable artifacts: a complete try_merge() fallback function, the statusCheckRollup jq query, and a full per-PR decision loop with categorical return codes. | 3 / 3 |
Workflow Clarity | A clear Discover → Confirm → Apply sequence with explicit validation checkpoints for a destructive batch op: scope confirmation with checklist, archived-repo pre-check, continue-on-failure, and a merge-strategy retry loop. | 3 / 3 |
Progressive Disclosure | A single self-contained SKILL.md with no bundle files and well-organized sections (decision tree, fallback chain, three-phase workflow, pitfalls, worked example); no nested references and no content that needs splitting out. | 3 / 3 |
Total | 11 / 12 Passed |