Content
100%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 lean, fully actionable policy skill: every rule is stated as a concrete command-level instruction, risky operations are gated by explicit pre-flight checks and confirmation steps, and the single explanatory paragraph supplies genuinely non-obvious repo context. Nothing is padded and nothing is missing for its scope.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body explains nothing Claude already knows; every line is repo-specific policy, and the one explanatory block (the shared-checkout "Why" with the NAS/server-shell example and the failed-push failure mode) is non-inferable context that earns its tokens — matching "every token earns its place" rather than level 4's "could be trimmed". | 5 / 5 |
Actionability | Guidance is fully executable with verbatim commands — `git checkout -b <new-branch>`, `git branch <new-branch>`, `git push` targeting `origin next_release`, `git status`, `git branch --show-current` — and covers the common cases of this workflow. | 5 / 5 |
Workflow Clarity | Validation checkpoints are explicit ("Run `git status` and `git branch --show-current` first, and don't assume the branch you last left the repo on is still checked out") and every push carries a human confirmation gate ("Ask for explicit confirmation immediately before running `git push`"), so the destructive-operation cap does not apply; the single-purpose action is unambiguous, satisfying the simple-skill exception. | 5 / 5 |
Progressive Disclosure | The skill is under 50 lines with no external references needed, and its four clearly-labeled sections (never branch without asking, default `next_release` flow, confirm before push, pre-flight checks) are well-organized, which the guidelines state is sufficient for a 5. | 5 / 5 |
Total | 20 / 20 Passed |