Content
21%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is a massive collection of GitHub Actions workflow templates that would be better served as a catalog of separate files with a concise overview. It is severely over-length, explains many patterns Claude already knows how to produce, contains several incomplete/pseudocode implementations (undefined functions like `callAI`, `analyzeIssue`), and lacks validation checkpoints for destructive operations like force-push rebasing and branch deletion.
Suggestions
Reduce the SKILL.md to a concise overview (under 80 lines) listing each automation pattern with a 1-2 sentence description, and move each complete workflow YAML into its own bundle file (e.g., `workflows/ai-review.yml`, `workflows/issue-triage.yml`).
Remove content Claude already knows how to generate: CODEOWNERS syntax, branch protection API calls, stale issue management with actions/stale, and generic best practices checklists.
Replace undefined function references (`callAI()`, `analyzeIssue()`, `ai.analyze()`) with actual executable implementations or remove them entirely.
Add explicit validation/verification steps for destructive operations: the auto-rebase workflow should verify the rebase succeeded before force-pushing, and the branch cleanup should include a dry-run or confirmation step.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | This is extremely verbose at 500+ lines, containing massive YAML blocks, multiple complete workflow files, and extensive boilerplate that Claude already knows how to produce. The 'When to Use This Skill' section, review comment templates, CODEOWNERS examples, and branch protection setup are all things Claude can generate without instruction. Very little here is novel knowledge that earns its token cost. | 1 / 5 |
Actionability | The YAML workflows and code snippets are mostly concrete and copy-pasteable, but several critical pieces are incomplete or pseudocode-like: `callAI(prompt)` is undefined, `analyzeIssue()` is referenced but never implemented, `ai.analyze()` and `ai.chat()` are undefined abstractions. The smart cherry-pick TypeScript is illustrative rather than executable. | 3 / 5 |
Workflow Clarity | While individual workflow files have implicit sequencing via GitHub Actions job ordering, there is no overarching guidance on how to adopt these patterns, no validation checkpoints for verifying workflows work correctly, and no error recovery guidance. The deployment workflow has an AI risk assessment but the `callAI` function is undefined, making the validation step non-functional. Destructive operations like force-push rebasing and branch cleanup lack verification steps. | 2 / 5 |
Progressive Disclosure | This is a monolithic wall of content with no bundle files to offload detail into. All 500+ lines are inlined in a single file with no references to supporting documents. The individual workflow files (ai-review.yml, issue-triage.yml, etc.) should clearly be separate files referenced from a concise overview, but instead everything is dumped inline. | 1 / 5 |
Total | 7 / 20 Passed |