Content
46%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 workflow is structured around concrete tool calls, but quality is dragged down by duplicated content, missing validation checkpoints for a batch operation, and dangling references to bundle files that do not exist. Fixing the references and adding verification steps would materially improve the score.
Suggestions
Remove the duplicated "可用资源" section and add the missing references/weekly-template.md and references/classification-rules.md files (or inline their key content).
Add validation checkpoints after the PR search (handle zero results) and after classification/report generation (review the output before presenting).
Tighten or justify the "IF-THEN 规则" note, which currently adds noise without actionable guidance.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly lean with named tools and a repo list, but the "可用资源" references section is duplicated verbatim and an unexplained "IF-THEN 规则" note adds noise, so it could be tightened. | 3 / 5 |
Actionability | Tool names, fields, and repos are concrete and executable, but the critical template and classification rules are only referenced (weekly-template.md, classification-rules.md) and those files do not exist in the bundle, leaving key details missing. | 3 / 5 |
Workflow Clarity | Four steps are clearly sequenced, but this batch operation (searching PRs across 5 repos and classifying them) has no validation checkpoints — no checks for empty results, login retrieval, or report review — so workflow clarity is capped at 3 per the batch-operation guideline. | 3 / 5 |
Progressive Disclosure | The body signals references via skill:// URIs, but the referenced files (classification-rules.md, weekly-template.md) do not exist in any bundle directory, so navigation points nowhere; the references block is also duplicated. | 2 / 5 |
Total | 11 / 20 Passed |