メニュー
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.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
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.