菜单
Load this skill when the user asks to review code, a PR, a diff, a changelist (CL), or requests a code review. Trigger words include: review, PR review, pull request review, diff review, code review, CL review, LGTM, look at this code.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when writing or reviewing REST controllers, HTTP endpoints, URL paths, route naming, HTTP verbs, status codes in a Spring Boot project
Use when writing, reviewing, or asking about Java code in a Spring Boot project — error handling, exceptions, catch blocks, constructor injection, @Autowired on fields, service methods, null returns, Optional
Use this skill when an AI agent needs to understand a distributed system at runtime. Covers: service dependency mapping, latency analysis per hop, change impact assessment, active incident investigation, SLO compliance, and error analysis. Invoke this skill for questions like: - "what does service X depend on?" - "where is the latency in this request chain?" - "what did this deploy break?" - "what is currently alerting?" - "is it safe to deploy right now?" - "who calls this service?" Do NOT use for: code review, writing code, static analysis, git operations.
| name | google-code-review |
| description | Load this skill when the user asks to review code, a PR, a diff, a changelist (CL), or requests a code review. Trigger words include: review, PR review, pull request review, diff review, code review, CL review, LGTM, look at this code. |
This skill implements the Google Engineering Code Review standard as documented at https://google.github.io/eng-practices/review/reviewer/. It provides a structured framework for evaluating code changes across seven distinct angles.
Each angle file contains concrete, actionable rules derived from the Google Engineering practices. The seven angles are:
| Angle | File | Covers |
|---|---|---|
| Design | angles/design.md | Architecture, system integration, placement of code |
| Functionality | angles/functionality.md | Correctness, edge cases, concurrency, user impact |
| Complexity | angles/complexity.md | Readability, over-engineering, function/class size |
| Tests | angles/tests.md | Test presence, correctness, quality, and simplicity |
| Naming, Comments & Style | angles/naming-comments-style.md | Identifiers, inline comments, style guide adherence |
| Documentation | angles/documentation.md | READMEs, guides, deprecation notices |
| Context & Health | angles/context-health.md | Whole-file context, code health trajectory |
When this skill is invoked by a subagent, the subagent's prompt will specify WHICH angle file to focus on. Load and apply ONLY that angle's rules — do not apply rules from other angle files. This keeps each reviewer's output focused and non-overlapping.
Every finding must include all three of these fields:
Blocking, Nit, or Positive
Blocking — must be resolved before approvalNit — optional polish, author may ignorePositive — acknowledge something done wellEnd every angle review with a brief ## Angle Assessment paragraph (2–4
sentences) summarizing what you found and whether this angle raises any
blockers.