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。
该技能材料显示其主要是用于代码审查的提示式工作流,且系统检查项已标注为 prompt-only、无密钥、无远程端点,整体风险较低。需注意其描述中包含读取 git diff、分派子代理等操作意图,但基于现有材料未见明确外发、提权或闭源失审等高风险红旗。
材料明确写明无需密钥或环境变量,未见要求访问 API token、云凭证或账户秘密,也未见诱导收集凭证的描述,凭证暴露面较低。
系统检查项标明无远程端点;文档中的网络相关内容仅涉及对 git 远端如 origin/main 的常规仓库同步语义,未声明将用户代码或数据发送到第三方服务,未见不明外发目标。
README 明确包含 git fetch、git log、git diff 等命令式步骤,并提到分派 Breaker/API Analyst 子代理;这表明其设计上可能触发本地审查流程或任务编排。此类本地命令/代理编排属于技能常规能力,材料中未见申请超出代码审查所需的系统权限。
该技能面向代码评审,按描述会读取分支差异、提交记录以及在 Deep 模式下读取完整变更文件;这意味着会访问本地仓库内容,但访问范围与声明用途基本一致,未见要求读取无关目录、系统敏感文件或额外资源。
来源为 GitHub 且标注 open-source,源码可审计,这是明显的降风险因素;但给出的仓库链接与技能名称/用途之间对应关系不够清晰,许可证未声明、star 为 0、维护状态未知,因此供应链可信度尚可但仍需核验具体仓库内容与维护情况。
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 "review" 技能: 1. 下载 https://raw.githubusercontent.com/microsoft/FluidFramework/main/.claude/skills/review/SKILL.md 2. 保存为 ~/.claude/skills/review/SKILL.md 3. 装好后重载技能,告诉我可以用了
请对这个功能分支做一次完整代码审查,重点检查逻辑正确性、API 兼容性、架构合理性、测试覆盖、性能风险和安全隐患,并按严重程度给出修改建议。
一份结构化审查报告,列出发现的问题、影响范围、严重级别及具体修改建议。
请审查这次接口改动,判断是否破坏向后兼容,是否符合现有命名与约定,并指出调用方可能受影响的地方。
关于兼容性与规范性的审查意见,包含潜在破坏性变更和改进方案。
在合并前帮我评审这组代码变更,找出可能遗漏的测试、性能瓶颈、重复实现以及安全问题,并给出是否建议合并的结论。
一份合并前评审结论,包含风险摘要、待修复项和是否建议合并。
/review # review current branch vs main
/review my-feature-branch # review a specific branch vs main
Spawns dedicated Breaker (correctness) and API Analyst (compatibility/conventions) sub-agents in parallel while the orchestrator performs the Inspector pass (architecture, tests, performance, security). Depth is user-selected.
Optimize for high-confidence, concise findings. Silence is better than speculation.
Target: $ARGUMENTS
Parse $ARGUMENTS:
my-feature-branch) -> diff that branch vs mainBefore doing anything, ask the user:
I can run a code review on your branch. Pick a depth (fastest to slowest):
- Skip — skip the review
- Quick — single-pass by orchestrator, all areas, no sub-agents
- Standard — full swarm: Breaker + API Analyst sub-agents + Inspector
- Deep — Standard + reads full changed files (not just diffs) for deeper analysis
Wait for the user's response. If they say skip, stop here.
<required> Immediately after the user picks a mode, create one task per applicable step using TaskCreate — before doing any other work. Mark each task in_progress when you start it and completed when you finish.Tasks to create by mode:
git fetch origin main && git log origin/main..HEAD --oneline && git diff --stat origin/main...HEAD
If a branch name was provided as argument, replace HEAD with origin/<branch-name> and add it to the fetch:
git fetch origin main <branch-name> && git log origin/main..origin/<branch-name> --oneline && git diff --stat origin/main...origin/<branch-name>
If on main and no branch name was provided, ask the user which branch to review.
Store the commit log, file list from --stat, and total $LINES_CHANGED.
Empty diff gate: Zero changed files -> report "No changes to review" and stop.
Size gate: >10,000 lines changed -> ask user to narrow scope before proceeding.
Exclude non-reviewable files from the file list: type declarations (.d.ts), lockfiles (pnpm-lock.yaml, package-lock.json), images, fonts, binaries, .map files, and generated API report files (*.api.md).
Read per-file diffs in batches (~50 files or ~500 changed lines per batch, whichever is smaller):
git diff origin/main...HEAD -- <file1> <file2> ...
For named branches, use origin/main...origin/<branch>.
After reading diffs, identify files where fuller context is needed — typically where the change touches a function that references shared state, calls other functions in the same file, or has fragmented hunks. Read those files in one batch using the Read tool.
Read every changed file in full (the version on the review branch). Use git show HEAD:<file> or git show origin/<branch>:<file> as appropriate.
If a change threads a new prop, callback, flag, or data field through a shared component or helper, read every changed call site and adjacent wrapper that accepts or forwards it.
…
在代码改动完成后,快速整理并更新对应技术文档与说明内容。
为 Copilot 创建的 PR 自动触发 ADO CI 流水线运行与验证。
帮助你严谨评估代码评审意见,澄清疑点后再决定是否采纳与实现
通过先写失败测试再实现代码,帮助稳定完成功能开发与缺陷修复。
在编写或修改测试时识别反模式,避免错误 mock 与污染生产代码。
用四阶段系统化排查框架定位缺陷根因,再制定可靠修复方案。
从正确性、测试、安全与性能等维度进行深入代码审查并给出改进建议
对本地或PR分支执行结构化代码审查,帮助提交或发布前发现问题。
在继续开发前发起代码审查,依据计划或需求检查实现质量与偏差。
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
模拟多角色工程师协作审查代码,快速发现质量、风险与设计问题。
帮助开发者进行智能代码审查、重构优化并自动生成测试用例。