Content
80%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.
Highly actionable and well-structured with executable commands and an exact output template. The main weakness is the absence of validation/verification for a batch PR-processing workflow, which caps workflow clarity.
Suggestions
Add a validation checkpoint after the fetch step (e.g. verify the gh command succeeded and warn if zero PRs are returned) to satisfy the batch-operation workflow-clarity requirement.
Include error-recovery guidance for the classify step, such as confirming every PR was assigned a module or explicitly placed under 其他.
Trim the trailing "Rules" items that restate the format template (group by module, - #number line) to remove minor redundancy.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean with no concept over-explanation and uses a tight table plus an exact format template, though the trailing "Rules" list mildly restates the format example (e.g. group by module, - #number lines). | 4 / 5 |
Actionability | It gives a copy-paste-ready `gh pr list` command with full JSON fields and a Linux variant, a concrete module-classification table, and an exact output template covering the common cases. | 5 / 5 |
Workflow Clarity | Steps 1–4 are clearly sequenced, but this batch operation (up to 100 PRs) lacks validation checkpoints — no check that the fetch succeeded, no handling of partial/empty results — so the batch-operation cap applies. | 3 / 5 |
Progressive Disclosure | It is self-contained with well-organized Step 1–4 sections and no need for external references, which is appropriate for a focused single-task skill. | 5 / 5 |
Total | 17 / 20 Passed |