メニュー
Use when running CodeRabbit CLI code reviews, reviewing uncommitted or committed changes, comparing branches, or authenticating with CodeRabbit. Triggers on "code review", "cr review", "coderabbit", "review my changes", "review this branch".
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
| name | coderabbit |
| description | Use when running CodeRabbit CLI code reviews, reviewing uncommitted or committed changes, comparing branches, or authenticating with CodeRabbit. Triggers on "code review", "cr review", "coderabbit", "review my changes", "review this branch". |
| disable-model-invocation | true |
Required baseline: CodeRabbit feedback must be interpreted through code-review-release/subskills/thermo-nuclear-code-quality-review/SKILL.md first. Whether CodeRabbit focuses on bugs, security, or style, the main agent must still check for ambitious structural simplification/code-judo opportunities, file-size blowups, spaghetti branching, boundary/type cleanliness, canonical helper reuse, and unnecessary abstraction/cast churn.
AI code review from the terminal. The command is cr (alias for coderabbit).
origin/HEAD with main/master fallbackcr auth login # OAuth login (opens browser)
cr auth login --api-key <key> # API key auth (non-interactive)
cr auth status # Check auth state
cr auth org # Switch organization
cr auth logout # Log out
Reviews take 2-10+ minutes. Set tool timeout to at least 600000ms (10 min).
# All changes (committed + uncommitted) — the default
cr review
# Uncommitted changes only (staged + unstaged)
cr review -t uncommitted
# Committed changes only
cr review -t committed
# Compare against a base branch
cr review --base <default-branch>
# Compare against a specific commit
cr review --base-commit abc123
# Feed project instructions to the reviewer
cr review -c .claude/CLAUDE.md -c .coderabbit.yaml
When calling cr review from a tool or subagent:
--agent for structured JSON findings--plain) is human-readable textcr review --agent -t uncommitted
Ready-for-PR/finalize loop:
origin/HEAD with main/master fallback.cr review --agent --base <base>.| Task | Command |
|---|---|
| Review everything | cr review |
| Uncommitted only | cr review -t uncommitted |
| Committed only | cr review -t committed |
| Against default branch | cr review --base <default-branch> |
| Against a commit | cr review --base-commit <sha> |
| With context files | cr review -c file1 -c file2 |
| Agent output | cr review --agent |
| Interactive mode | cr review --interactive |
| Check auth | cr auth status |
| Login | cr auth login |
| Switch org | cr auth org |
| Update CLI | cr update |
Timeout too short. Default 2-minute timeouts kill the review mid-flight. Always set 10+ minutes (600000ms).
Wrong review type. cr review includes both committed and uncommitted changes. Use -t uncommitted to review only working-tree changes.
No context files. Without -c, the reviewer lacks project conventions. Pass your CLAUDE.md or coderabbit.yaml for better results.
Treating CodeRabbit as sufficient by itself. CodeRabbit output is one review input, not the approval bar. Always apply the thermo-nuclear code-quality baseline before declaring the branch clean.
Use when managing per-project backlog folders, feature queues, async agent work items, lifecycle status updates, PR/done/blocked transitions, or status reviews inside an Obsidian vault.
Use when prototyping, authoring agent skills, or asking the agent to zoom out and map unfamiliar code before acting.
Create new agent skills with proper structure, progressive disclosure, and bundled resources. Use when user wants to create, write, or build a new skill.
Use when reviewing architecture boundaries, hexagonal architecture, clean architecture layers, dependency direction, trust boundaries, uploads, parsers, redirects, or web security risks.
Use when configuring or running hexcheck in a Go repo for hexagonal architecture boundaries, adapter business-logic warnings, and mock discipline.
Find deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI-navigable.