Content
88%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 content is a strong, highly actionable workflow: every step carries executable commands, explicit validation checkpoints, and error-recovery guidance, with dependencies on external materials clearly signaled. The only weaknesses are minor — a few justificatory sentences that could be trimmed and no progressive-disclosure structure beyond the inline body.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is efficient and assumes Claude's competence, but a few rationale sentences could be trimmed, e.g., "Before jumping to a fix, verify assumptions. This step prevents wasted effort on the wrong solution" and the 7b explanation "This is valuable when you cannot fully predict how the external system handles the input... It proves the test is meaningful and not a no-op." | 4 / 5 |
Actionability | Fully executable guidance throughout: the exact `gh issue view ISSUE --json ...` command, a branch-name template with a concrete example (`fix/3529-dedup-sort-escape`), `/worktree-from-issue {branch-name}`, `make format && make lint`, and a copy-paste report template covering the common cases. | 5 / 5 |
Workflow Clarity | Eight clearly numbered steps with explicit validation checkpoints and feedback loops: cwd verification ("Stop with an error if the cwd cannot be set"), plan approval via ExitPlanMode, "Stop and fix issues before proceeding to the next step", the revert-fix/run-test/restore regression-proof loop, and "Wait for the human to confirm before committing." | 5 / 5 |
Progressive Disclosure | The ~200-line body is well-organized into clearly signaled sections and points one level deep to external materials (`/worktree-from-issue`, `@CLAUDE.md`, `@.claude/rules/worktree.md`); no bundle files exist, and the rubric reserves a 5-by-organization alone for skills under 50 lines. | 4 / 5 |
Total | 18 / 20 Passed |