Content
82%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 tight, highly actionable orchestration skill: verbatim commands, fully specified agent prompts, and a concrete output format with a clear verdict taxonomy. The main gaps are missing validation/error-recovery checkpoints in the workflow and a slightly long inline agent-task spec that could be split out.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is efficient and command-first with no explanations of concepts Claude already knows — git commands, agent prompts, and output structure are given directly. Minor trims are possible (the opening paragraph repeats the frontmatter description; "(MANDATORY)" and the parenthetical /pr-iterate note are slight padding), so it sits just below the every-token-earns-its-place level-5 anchor. | 4 / 5 |
Actionability | Fully executable throughout: verbatim commands (`git fetch origin`, `git merge-base origin/main HEAD`, `git diff <merge-base>..HEAD --name-only`), complete copy-paste-ready agent task prompts with explicit overrides, a concrete report template with a table schema, and the exact closing question. This matches the fully-executable level-5 anchor. | 5 / 5 |
Workflow Clarity | A clear 0 → 1 → 2 sequence with concrete commands at each step and explicit guardrails ("Do NOT rebase or modify the current branch", "Do NOT post anything to GitHub", "Skip steps 6 and 7 entirely"). It falls short of level 5 because there are no explicit validation checkpoints or error-recovery loops — e.g., what to do if the merge-base lookup fails, the branch has no unique commits, or an agent returns ambiguous output. | 4 / 5 |
Progressive Disclosure | The skill has no bundle files, and the body is well-organized into numbered protocol steps with the detailed agent-task specs appropriately inline as core instruction. File references are one level deep and clearly signaled (`.agents/skills/review-pr/SKILL.md`, `docs/adr/`, `docs/spec.html`, `docs/diary.md`). At ~75 lines it exceeds the under-50-line simple-skill case, and the Agent 2 prompt's six-step spec could arguably live in a reference file, so it does not quite reach a 5. | 4 / 5 |
Total | 17 / 20 Passed |