调用多模型交叉审查代码变更、PR与高风险修改,辅助发现缺陷与争议点
复制安装指令,让 AI 自动完成配置 · 推荐新手
请帮我安装 askskill 上的 "council-review" 技能: 1. 下载 https://raw.githubusercontent.com/microsoft/vscode-team-kit/main/model-council/skills/council-review/SKILL.md 2. 保存为 ~/.claude/skills/council-review/SKILL.md 3. 装好后重载技能,告诉我可以用了
请对这个 PR 做多模型交叉审查,重点检查逻辑错误、边界条件、可维护性和潜在回归风险,并汇总一致意见与分歧点。
一份综合审查结果,包含主要问题、风险等级、修复建议,以及不同模型意见的共识与分歧。
请让多个模型独立检查最近这批代码改动,找出可能引入的 bug、遗漏的测试场景和不安全实现,并按严重程度排序。
按严重程度排序的问题清单,附带触发条件、影响范围和建议补充的测试点。
请组织多个模型对同一组审查发现进行讨论,比较哪些问题成立、哪些结论有争议,并给出最终建议是否可以合并。
一份讨论后的最终结论,说明应修复的问题、可接受的风险,以及是否建议合并。
Read-only review powered by a council of model-pinned reviewers. Changes are usually not 100% correct — this skill exists to catch what slipped through. The goal is not broad commentary; it is a short list of concrete issues where the council agrees, plus transparent disclosure of where they disagree.
Spawn subagents with the model parameter to pin each to a different model. Use all three when available, at least two otherwise:
GPT-5.5Claude Opus 4.6GPT-5.3-CodexEach subagent receives the same system preamble:
You are an independent subagent for read-only code research. MUST stay read-only. Stay within the requested scope. Do not speculate. Do not suggest patches. For every finding, cite exact files and lines. Form your own view from first principles. Do not anchor to the provided context.
Before fanning out, build a preliminary orientation that each reviewer will receive. This is a starting point, not ground truth — reviewers are expected to contradict it if their own investigation leads elsewhere. Do paste the raw diff — reviewers have tools to read code themselves.
The summary should include only:
Keep the summary under ~50 lines. Do not flag risk areas or pre-diagnose issues — leave reviewers to form independent assessments. They get better results reading code in context than scanning a pre-interpreted summary.
Spawn all subagents in parallel at once:
model parameter. Prepend the system preamble to the reviewer prompt below.Classify every finding into one of three buckets:
Drop style chatter, linter-catchable nits, and low-confidence speculation entirely.
Trigger this phase when the user asks to "discuss", "debate", or "cross-review" the findings, or when Phase 4 produces contested findings that could benefit from a second look.
This deliberation step is lightweight — it only re-examines the synthesized findings, not the full codebase.
…
帮助你快速检索 GitHub 中分配给你、待分诊或自定义条件的议题与 PR。
用于代码与分支审查,综合检查正确性、兼容性、架构、测试、性能与安全问题。