review-as-json
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseReview (JSON output)
审查(JSON输出)
Run the skill on this change — follow it exactly: Step 0 (find the project's reference material), all six passes in order, then the verdict.
reviewInstead of a prose report, emit the result as a single JSON object:
json
{
"verdict": "approve | changes_needed | hold",
"summary": "One or two sentences on the change overall.",
"findings": [
{
"severity": "critical | warning | info",
"type": "issue | suggestion | question",
"category": "mission | architecture | implementation | craft | security | performance",
"title": "Short imperative, e.g. 'Sanitize user input before query'",
"description": "What it is, why it matters, and the consequence.",
"file": "path/to/file.ts",
"line": 42,
"suggestedFix": "Optional concrete fix."
}
],
"passes": {
"mission": "met | missing | different | unclear",
"architecture": "ok | concern | blocking",
"implementation": "ok | concern | blocking",
"craft": "ok | concern | blocking",
"security": "ok | concern | blocking",
"performance": "ok | concern | blocking"
},
"sectionSummaries": {
"mustFix": "1-2 sentences on the critical findings, if any.",
"shouldFix": "1-2 sentences on the warnings, if any.",
"questions": "1-2 sentences on the questions, if any.",
"suggestions": "1-2 sentences on the suggestions, if any."
}
}Rules for the output:
- Emit only the JSON object. No preamble, no commentary, no markdown fences around it.
- Severity. blocks merge (bugs, security holes, data loss).
criticalshould be addressed but isn't a blocker.warningis a note.info - Type. is something you need answered rather than a defect.
questionis an optional improvement. Everything else is ansuggestion.issue - Category. Every finding's is the pass that found it — this is how the passes show up in CI output.
category - Verify before you flag still applies. Every finding needs a concrete file, line, and consequence. If you can't verify it, make it a .
question - If there are no findings, emit an empty array and an
findingsverdict.approve
针对本次变更运行 skill——严格遵循以下步骤:步骤0(查找项目参考资料),按顺序完成全部六轮检查,然后给出结论。
review不生成散文式报告,而是将结果输出为单个JSON对象:
json
{
"verdict": "approve | changes_needed | hold",
"summary": "关于本次变更的总体概述,1-2句话即可。",
"findings": [
{
"severity": "critical | warning | info",
"type": "issue | suggestion | question",
"category": "mission | architecture | implementation | craft | security | performance",
"title": "简短祈使句,例如'查询前清理用户输入'",
"description": "问题内容、影响原因及后果。",
"file": "path/to/file.ts",
"line": 42,
"suggestedFix": "可选的具体修复方案。"
}
],
"passes": {
"mission": "met | missing | different | unclear",
"architecture": "ok | concern | blocking",
"implementation": "ok | concern | blocking",
"craft": "ok | concern | blocking",
"security": "ok | concern | blocking",
"performance": "ok | concern | blocking"
},
"sectionSummaries": {
"mustFix": "若存在严重问题,用1-2句话概述。",
"shouldFix": "若存在警告,用1-2句话概述。",
"questions": "若存在疑问,用1-2句话概述。",
"suggestions": "若存在建议,用1-2句话概述。"
}
}输出规则:
- 仅输出JSON对象。不要前置说明、评论,也不要用markdown围栏包裹。
- 严重程度:会阻止合并(如bug、安全漏洞、数据丢失问题)。
critical应被处理但不阻止合并。warning仅为提示信息。info - 类型:是需要解答的问题而非缺陷。
question是可选的改进建议。其余情况均为suggestion。issue - 分类:每个问题的对应发现该问题的检查环节——这是检查结果在CI输出中的呈现方式。
category - 仍需遵循“先验证再标记”原则。每个问题都需要明确的文件路径、行号和影响后果。若无法验证,则标记为。
question - 若无问题,输出空的数组,并给出
findings结论。approve