メニュー
Review local code changes before commit, surface bugs and security issues, and feed findings back into an AI coding loop. Use when an agent needs to run terminal-native code review on a Git repo and act on the results.
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 CLI |
| description | Review local code changes before commit, surface bugs and security issues, and feed findings back into an AI coding loop. Use when an agent needs to run terminal-native code review on a Git repo and act on the results. |
AI code reviews directly in the terminal before commit. Reviews uncommitted changes, supports agent-friendly output modes, and integrates well with coding-agent fix loops.
# Install script
curl -fsSL https://cli.coderabbit.ai/install.sh | sh
# Homebrew
brew install coderabbit
The CLI binary is coderabbit, with cr as a short alias.
# Browser-based login
cr auth login
# API key login
coderabbit auth login --api-key "cr-************"
CodeRabbit links the CLI to your account so reviews can use plan limits, team learnings, and codebase context where available.
# Review current repo changes
cr
# Plain text output
cr --plain
# Minimal output for AI coding agents
cr --prompt-only
# Review against a non-main base branch
cr --base develop
# Explicitly review uncommitted changes
cr --prompt-only --type uncommitted
The CLI must be run from inside a Git repository.
cr --prompt-only for minimal agent-oriented output--plain for detailed terminal review output without the full interactive UIImplement the task, then run cr --prompt-only --type uncommitted in the background.
Fix major issues only. Run one final pass after the fixes and stop after two loops.
Use CodeRabbit to detect issues, then use the coding agent to implement the fixes. This is especially useful for race conditions, memory leaks, logic bugs, and security findings that are easier to address before commit.
Render videos, audio, and still images programmatically with React using terminal commands. Use when an agent needs to preview compositions, list available compositions, or render media artifacts from a Remotion project.
Web search, page extraction, site crawling, and multi-source research from the terminal. Use when an agent needs to search the web, extract content from URLs, crawl sites with an objective, or run deep research across sources.
Manage Sentry releases, upload debug symbols and source maps, send events, monitor crons, and stream logs. Use when an agent needs to create releases, upload source maps, manage debug info files, or send test events to Sentry.
Search GIFs from Tenor and Giphy via CLI or TUI, download GIFs, and extract stills or sprite sheets. Use when an agent needs to find a GIF by keyword, download it, or generate a still frame or contact sheet from a GIF.
Manage AWS services — EC2, S3, Lambda, IAM, and more. Use when an agent needs to provision cloud infrastructure, manage S3 buckets, invoke Lambda functions, or configure AWS resources.
Search, authenticate, and execute tools across 1000+ apps from the terminal. Use when an agent needs to discover available integrations, connect to third-party services, or execute actions across SaaS apps.