Content
77%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.
The body is a well-structured, lean workflow with an explicit review-until-clean feedback loop and good sequencing. Its main weakness is actionability: the steps tell Claude what to do but omit the exact mechanism for summoning and checking the Copilot review.
Suggestions
Specify how to summon the Copilot review (e.g. the exact comment, command, or UI action) and how to detect it has returned, so the polling step is fully executable.
Optionally give the concrete branch/PR commands (e.g. `gh pr create`, `gh pr review`) so the flow is copy-paste ready rather than left to inference.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean, assumes Claude's competence, and avoids explaining PRs or reviews from scratch; only a few soft sentences ('When in doubt, treat it as a real change...') could be trimmed. | 4 / 5 |
Actionability | Steps are concrete (branch/commit/open PR, poll ~10s, address comments, push, re-summon, merge) but key mechanisms are missing — it never says how to 'Summon a Copilot code review' or how to check whether it has landed. | 3 / 5 |
Workflow Clarity | A clearly sequenced 5-step process with an explicit validation feedback loop ('Repeat until the review is clean') and a polling checkpoint; this is not a destructive/batch op so no cap applies. | 5 / 5 |
Progressive Disclosure | Under 50 lines, single-purpose, and well-organized with a heading, prose intro, and numbered list; no external references are needed, satisfying the simple-skill exception. | 5 / 5 |
Total | 17 / 20 Passed |