Skip to content
Back to search

coderabbit

CodeRabbit is the governance and control layer for agentic software development helping developers trust AI written code.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today

Check it deeper on your machine

It has changed since its scan on 13 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/coderabbitai/skills with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/coderabbitai/skills'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/coderabbitai--skills/badge.svg)](/listings/coderabbitai--skills)

HTML

<a href="/listings/coderabbitai--skills"><img src="/api/listings/coderabbitai--skills/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

**Claude MCP plugin:** the repository root on `main` is the installable Claude 1.3.0 candidate. It connects remote CodeRabbit tools through MCP and uses the standalone CodeRabbit CLI for local reviews. It has no automatic review hook. Run `/coderabbit:setup` to check both: **connect MCP → sign in to the CLI if needed → “Review my changes.”** The two sign-ins are separate; an existing CLI installation and login can be reused. See [distribution status](DISTRIBUTION_CHANNELS.md) for directory publication, which is separate from source availability.

The remote connection is declared only in Claude's manifest. Shared skills and other host plugins keep their standalone CLI workflow. Local reviews do not reserve MCP reviews or pass review handoffs through shell commands. A hosted connector alone cannot collect local files.

For setup, recovery and data handling, read the [MCP setup guide](docs/mcp.md). Independently assembled MCP hook/direct packages remain development candidates with their own qualification requirements; see [packaging](packaging/README.md).

**Released CLI integration:** install the CodeRabbit CLI via the [CLI docs](https://docs.coderabbit.ai/cli), then tell your agent: **“Review my code.”** The review command starts the authentication flow when needed.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/plugin.json · Manifest · 35 lines
  1. {
  2. "name": "coderabbit",
  3. "displayName": "CodeRabbit",
  4. "version": "1.3.0",
  5. "description": "CodeRabbit is the governance and control layer for agentic software development helping developers trust AI written code.",
  6. "homepage": "https://github.com/coderabbitai/skills/blob/main/docs/mcp.md",
  7. "author": {
  8. "name": "CodeRabbit",
  9. "email": "support@coderabbit.ai",
  10. "url": "https://coderabbit.ai"
  11. },
  12. "repository": "https://github.com/coderabbitai/skills",
  13. "license": "MIT",
  14. "keywords": [
  15. "code-review",
  16. "autofix",
  17. "ai",
  18. "coderabbit",
  19. "claude-code"
  20. ],
  21. "documentationUrl": "https://docs.coderabbit.ai/api/platform-api#use-hosted-mcp",
  22. "supportUrl": "https://www.coderabbit.ai/contact-us/support",
  23. "privacyPolicyUrl": "https://www.coderabbit.ai/privacy-policy",
  24. "termsOfServiceUrl": "https://www.coderabbit.ai/terms-of-service",
  25. "icon": "./assets/coderabbit-small.png",
  26. "mcpServers": {
  27. "coderabbit": {
  28. "type": "http",
  29. "url": "https://mcp.coderabbit.ai/",
  30. "oauth": {
  31. "scopes": "connection account:read usage:read pull_requests:read review"
  32. }
  33. }
  34. }
  35. }
README.md · README · 374 lines
  1. # CodeRabbit Skills and Plugins
  2. [![Standalone release](https://img.shields.io/badge/standalone-1.2.2-blue)](https://github.com/coderabbitai/skills/releases/tag/v1.2.2)
  3. [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
  4. [![Agents](https://img.shields.io/badge/works_with-35%2B_agents-brightgreen)](#supported-agents)
  5. The canonical home for CodeRabbit's agent-native skills and plugin packaging.
  6. Use it to install AI-powered code review into 35+ coding agents, Gemini CLI,
  7. Antigravity CLI, Claude Code, Cursor, and other supported agent environments.
  8. CodeRabbit detects bugs, security issues, and quality risks before you merge.
  9. ## Quickstart
  10. **Claude MCP plugin:** the repository root on `main` is the installable Claude
  11. 1.3.0 candidate. It connects remote CodeRabbit tools through MCP and uses the
  12. standalone CodeRabbit CLI for local reviews. It has no automatic review hook.
  13. Run `/coderabbit:setup` to check both: **connect MCP → sign in to the CLI if
  14. needed → “Review my changes.”** The two sign-ins are separate; an existing CLI
  15. installation and login can be reused. See [distribution status](DISTRIBUTION_CHANNELS.md)
  16. for directory publication, which is separate from source availability.
  17. The remote connection is declared only in Claude's manifest. Shared skills and
  18. other host plugins keep their standalone CLI workflow. Local reviews do not
  19. reserve MCP reviews or pass review handoffs through shell commands. A hosted
  20. connector alone cannot collect local files.
  21. For setup, recovery and data handling, read the [MCP setup guide](docs/mcp.md).
  22. Independently assembled MCP hook/direct packages remain development candidates
  23. with their own qualification requirements; see [packaging](packaging/README.md).
  24. **Released CLI integration:** install the CodeRabbit CLI via the [CLI docs](https://docs.coderabbit.ai/cli),
  25. then tell your agent: **“Review my code.”** The review command starts the
  26. authentication flow when needed.
  27. ## Installation
  28. The standalone integration paths below remain available. The native Claude
  29. plugin on `main` is the MCP candidate described above; directory and official
  30. marketplace publication require their own approval and catalog updates.
  31. ### 1. Install the CodeRabbit CLI
  32. Use the [CLI docs](https://docs.coderabbit.ai/cli) for the primary install path.
  33. They cover Homebrew, the install script, and CLI usage.
  34. ### 2. Install the agent integration
  35. Choose the path that matches your coding agent.
  36. #### Skills installer
  37. For agents that support portable `SKILL.md` files, use the
  38. [skills docs](https://docs.coderabbit.ai/cli/skills).
  39. ```bash
  40. npx skills add coderabbitai/skills
  41. ```
  42. Installation options for the skills installer:
  43. | Flag | Purpose |
  44. | -------------- | ------------------------------------------------ |
  45. | `-g, --global` | Install to user directory instead of project |
  46. | `-a, --agent` | Target specific agents (for example `claude-code`) |
  47. | `-s, --skill` | Install particular skills by name |
  48. | `--all` | Install all skills to all agents without prompts |
  49. #### Claude Plugin
  50. In Claude Code, install the plugin from the official marketplace:
  51. ```text
  52. /plugin marketplace update
  53. /plugin install coderabbit@claude-plugins-official
  54. ```
  55. CodeRabbit is also listed in the Claude directory. Add it from
  56. **Customize > Plugins > Discover** in
  57. claude.ai or the Claude desktop app, and it will sync to Claude Code sessions
  58. where you're signed in with the same claude.ai account. If you install from
  59. both places, Claude Code loads the marketplace copy.
  60. For the MCP candidate from this source, use `/coderabbit:setup` and the
  61. [MCP setup guide](docs/mcp.md). Published marketplace copies may still use the
  62. [Claude Code integration guide](https://docs.coderabbit.ai/cli/claude-code-integration).
  63. #### Cursor Plugin
  64. This repository now includes Cursor marketplace metadata in
  65. [`/.cursor-plugin/plugin.json`](.cursor-plugin/plugin.json).
  66. After publication, Cursor marketplace installs use:
  67. ```text
  68. /add-plugin coderabbit
  69. ```
  70. For the current recommended setup, see the
  71. [Cursor integration guide](https://docs.coderabbit.ai/cli/cursor-integration).
  72. #### Gemini CLI Extension
  73. Gemini CLI users can install the native extension from the published
  74. [standalone `v1.2.2` tag](https://github.com/coderabbitai/skills/releases/tag/v1.2.2):
  75. ```bash
  76. gemini extensions install https://github.com/coderabbitai/skills --ref v1.2.2
  77. gemini extensions list
  78. ```
  79. The repository-root [`gemini-extension.json`](gemini-extension.json) packages
  80. the portable skills, `/coderabbit:review` command, and both subagents.
  81. This native extension is separate from the generic skills-installer path above.
  82. #### Antigravity CLI Plugin
  83. Antigravity CLI users can install the native plugin directly from this
  84. repository:
  85. ```bash
  86. agy plugin install https://github.com/coderabbitai/skills
  87. agy plugin list
  88. ```
  89. The repository-root [`plugin.json`](plugin.json) manifest packages the skills,
  90. review command, and both subagents for Antigravity CLI's plugin discovery.
  91. #### Codex App
  92. Codex users can install the official CodeRabbit plugin by following the
  93. [Codex app integration guide](https://docs.coderabbit.ai/cli/codex-integration#codex-app).
  94. Native Codex metadata in [`.codex-plugin/plugin.json`](.codex-plugin/plugin.json)
  95. references the existing `skills/` directory. Marketplace publication remains a
  96. separate step; this source change does not update the live listing.
  97. For an at-a-glance inventory of active and repo-packaged distribution paths, see
  98. [DISTRIBUTION_CHANNELS.md](DISTRIBUTION_CHANNELS.md).
  99. ## What Lives Here
  100. | Path | Purpose |
  101. | --- | --- |
  102. | `skills/` | Portable CodeRabbit skills for agents that support `SKILL.md`. |
  103. | `.claude-plugin/` | Native Claude manifest with its remote MCP connection; no review hook. |
  104. | `commands/` | Native review commands for Gemini CLI and Antigravity CLI. |
  105. | `agents/` | `code-reviewer` and `autofix` subagents shipped to Claude Code, Gemini CLI, and Antigravity CLI. |
  106. | `.cursor-plugin/` | Cursor marketplace metadata. |
  107. | `.codex-plugin/` | Codex metadata referencing the shared skills. |
  108. | `gemini-extension.json` | Gemini CLI extension manifest. |
  109. | `plugin.json` | Antigravity CLI plugin manifest. |
  110. | `assets/` | Shared marketplace and brand assets. |
  111. | `DISTRIBUTION_CHANNELS.md` | Maintainer inventory of live, packaged, and in-development channels. |
  112. ## Usage
  113. Once installed, just ask your agent:
  114. ```text
  115. Review my code
  116. Check for security issues
  117. What's wrong with my changes?
  118. Run a code review
  119. Review my PR
  120. Review the directory at ../my-service
  121. ```
  122. For standalone installs, the agent will automatically:
  123. 1. Check if CodeRabbit CLI is installed
  124. 2. Run the review on your changes
  125. 3. Present findings grouped by severity
  126. 4. Optionally fix issues and re-review
  127. When you ask for a specific review directory, the agent can pass CodeRabbit CLI
  128. `--dir <path>` after confirming that path is inside an initialized Git working
  129. tree.
  130. ## Supported Agents
  131. CodeRabbit supports 35+ coding agents.
  132. | Agent | Project Path | Global Path |
  133. | ------------------ | ---------------------- | -------------------------------------- |
  134. | Amp, Kimi Code CLI | `.agents/skills/` | `~/.config/agents/skills/` |
  135. | Antigravity CLI | `.agents/skills/` | `~/.gemini/config/skills/` |
  136. | Claude Code | `.claude/skills/` | `~/.claude/skills/` |
  137. | Cline | `.cline/skills/` | `~/.cline/skills/` |
  138. | CodeBuddy | `.codebuddy/skills/` | `~/.codebuddy/skills/` |
  139. | Codex | `.codex/skills/` | `~/.codex/skills/` |
  140. | Command Code | `.commandcode/skills/` | `~/.commandcode/skills/` |
  141. | Continue | `.continue/skills/` | `~/.continue/skills/` |
  142. | Crush | `.crush/skills/` | `~/.config/crush/skills/` |
  143. | Cursor | `.cursor/skills/` | `~/.cursor/skills/` |
  144. | Droid | `.factory/skills/` | `~/.factory/skills/` |
  145. | Gemini CLI | `.gemini/skills/` | `~/.gemini/skills/` |
  146. | GitHub Copilot | `.github/skills/` | `~/.copilot/skills/` |
  147. | Goose | `.goose/skills/` | `~/.config/goose/skills/` |
  148. | Junie | `.junie/skills/` | `~/.junie/skills/` |
  149. | Kilo Code | `.kilocode/skills/` | `~/.kilocode/skills/` |
  150. | Kiro CLI | `.kiro/skills/` | `~/.kiro/skills/` |
  151. | Kode | `.kode/skills/` | `~/.kode/skills/` |
  152. | MCPJam | `.mcpjam/skills/` | `~/.mcpjam/skills/` |
  153. | Moltbot | `skills/` | `~/.moltbot/skills/` |
  154. | Mux | `.mux/skills/` | `~/.mux/skills/` |
  155. | Neovate | `.neovate/skills/` | `~/.neovate/skills/` |
  156. | OpenClaude IDE | `.openclaude/skills/` | `~/.openclaude/skills/` |
  157. | OpenCode | `.opencode/skills/` | `~/.config/opencode/skills/` |
  158. | OpenHands | `.openhands/skills/` | `~/.openhands/skills/` |
  159. | Pi | `.pi/skills/` | `~/.pi/agent/skills/` |
  160. | Pochi | `.pochi/skills/` | `~/.pochi/skills/` |
  161. | Qoder | `.qoder/skills/` | `~/.qoder/skills/` |
  162. | Qwen Code | `.qwen/skills/` | `~/.qwen/skills/` |
  163. | Replit | `.agent/skills/` | N/A (project-only) |
  164. | Roo Code | `.roo/skills/` | `~/.roo/skills/` |
  165. | Trae | `.trae/skills/` | `~/.trae/skills/` |
  166. | Trae CN | `.trae/skills/` | `~/.trae-cn/skills/` |
  167. | Windsurf | `.windsurf/skills/` | `~/.codeium/windsurf/skills/` |
  168. | Zencoder | `.zencoder/skills/` | `~/.zencoder/skills/` |
  169. ## Available Skills
  170. ### [code-review](skills/code-review/SKILL.md)
  171. AI-powered code review that finds bugs, security issues, and suggests improvements using CodeRabbit.
  172. **Use when:**
  173. - You want to review code changes before committing or merging
  174. - Checking for bugs, security vulnerabilities, or anti-patterns
  175. - Getting PR feedback or suggestions for improvements
  176. - Running automated code quality checks
  177. **Categories covered:** Bug detection, security analysis, code quality, performance issues, best practices
  178. **Triggers:** "review my code", "check for bugs", "security review", "PR feedback", "run coderabbit"
  179. **Capabilities:**
  180. - Analyzes code changes for bugs, security issues, and anti-patterns
  181. - Preserves finding severities (critical, major, minor, trivial, info, none)
  182. - Supports autonomous fix-review cycles
  183. - Reviews tracked changes by default, with committed and uncommitted scopes
  184. - Supports directory-scoped reviews through CodeRabbit CLI `--dir <path>`
  185. ### [autofix](skills/autofix/SKILL.md)
  186. Safe fix workflow for unresolved CodeRabbit GitHub PR review threads, with per-issue review and approval.
  187. **Use when:**
  188. - You already have an open GitHub PR reviewed by CodeRabbit
  189. - You want to apply suggested fixes from unresolved current CodeRabbit review threads
  190. - You want guided fixes with explicit approval for each change
  191. **Categories covered:** Review-thread extraction, issue prioritization, guarded fixes, consolidated commit and PR summary
  192. **Triggers:** "coderabbit autofix", "fix coderabbit", "cr fix"
  193. **Capabilities:**
  194. - Fetches unresolved current CodeRabbit review threads for the current PR
  195. - Parses and prioritizes issues by severity
  196. - Applies fixes only after validating the issue and getting approval
  197. - Produces a single consolidated commit and posts a PR summary comment
  198. ### [coderabbit-config](skills/coderabbit-config/SKILL.md)
  199. Create, update, or validate repository configuration through the official CLI.
  200. Ask “Set up CodeRabbit configuration”, “Make CodeRabbit reviews more thorough”,
  201. or “Validate .coderabbit.yaml”. Where supported, invoke `$coderabbit-config` or
  202. `/coderabbit-config`; the Claude plugin uses `/coderabbit:coderabbit-config`.
  203. The distinct name leaves the host's `/config` command available.
  204. Requires an official CLI supporting config protocol v2 (verified with 0.9.3).
  205. The CLI owns defaults, schema checks, previews, and writes. The skill supports
  206. review-style changes and the CLI's guided path settings; unsupported settings
  207. remain a CLI capability gap. Missing commands require an upgrade, never an
  208. agent-side YAML editor. Validation alone does not change files; saves require
  209. approval of the preview. Configuration does not require login or run a review.
  210. This skill is included in the source package and the published
  211. [standalone `v1.2.2` archive](https://github.com/coderabbitai/skills/releases/tag/v1.2.2).
  212. Marketplace caches require their normal publication update to deliver it.
  213. ## Plugin Components
  214. The native Claude candidate combines a remote MCP connection and setup command
  215. with standalone CLI review and configuration skills. It registers no hooks and
  216. supplies no managed review runtime. Independently assembled MCP variants have
  217. separate qualification requirements.
  218. ### Claude Code
  219. - Skills: `code-review`, `autofix`, and `coderabbit-config`. Run them as
  220. `/coderabbit:code-review`, `/coderabbit:autofix`, and
  221. `/coderabbit:coderabbit-config`, or ask in plain language.
  222. - Subagents: `code-reviewer` runs a CodeRabbit review in its own context, and
  223. `autofix` collects and checks unresolved CodeRabbit review threads
  224. and returns a fix plan without editing anything. Mention them as
  225. `@agent-coderabbit:code-reviewer` and `@agent-coderabbit:autofix`.
  226. - Manifest: `.claude-plugin/plugin.json`
  227. The Claude plugin adds `/coderabbit:setup` for MCP connection and standalone CLI readiness.
  228. For review, use `/coderabbit:code-review` instead of the earlier
  229. `/coderabbit:coderabbit-review`.
  230. Local review uses the user's installed and authenticated CodeRabbit CLI and
  231. requires host execution permission and access to the intended Git working tree.
  232. MCP sign-in does not authenticate the CLI. A host without local execution can
  233. use remote tools but cannot run a local review. No hook trust is needed.
  234. ### Cursor
  235. - Marketplace manifest: `.cursor-plugin/plugin.json`
  236. - Skills source: `skills/`
  237. ### Antigravity CLI
  238. For sandboxed CLI sessions that cannot find or execute an installed CodeRabbit
  239. CLI, follow the [scoped executable-access recovery](skills/code-review/references/antigravity-sandbox.md).
  240. It keeps diagnostics sandboxed and retains per-command host approval for auth and review.
  241. - Native plugin manifest: `plugin.json`
  242. - Skills source: `skills/`
  243. - Review command: `/coderabbit:review` from `commands/coderabbit/review.toml` (converted to a skill during installation)
  244. - Subagents: `agents/code-reviewer.md`, `agents/autofix.md`
  245. ### Gemini CLI
  246. - Native extension manifest: `gemini-extension.json`
  247. - Skills source: `skills/`
  248. - Slash command: `/coderabbit:review` (`commands/coderabbit/review.toml`)
  249. - Subagents: `agents/code-reviewer.md`, `agents/autofix.md`
  250. ## Data and Privacy
  251. The native Claude plugin contains three portable skills, a setup command, two
  252. subagents and a remote MCP connection. It declares no automatic hooks.
  253. - **Local commands**: reviews use the user's installed CodeRabbit CLI. The
  254. `autofix` skill uses GitHub CLI (`gh`) to read review threads and, when
  255. authorized, post or create a pull request. The `autofix` subagent only reads.
  256. - **Data sent**: local reviews send selected changes and required repository
  257. context to CodeRabbit. MCP tools exchange requested account and repository
  258. data with CodeRabbit; `autofix` exchanges pull-request data with GitHub.
  259. Setup alone does not authorize an upload or start a review.
  260. - **Credentials**: Claude manages connector OAuth. The standalone CLI and `gh`
  261. manage their own logins. The plugin does not read Claude's OAuth store or
  262. copy credentials between the connector and the CLI.
  263. - **Policy**: see the [CodeRabbit Privacy Policy](https://www.coderabbit.ai/privacy-policy)
  264. for handling and retention, or contact <support@coderabbit.ai>.
  265. The `coderabbit-config` skill runs local CLI configuration commands. The CLI
  266. fetches the public configuration schema over HTTPS and validates YAML locally;
  267. it does not submit repository code for review. Approved saves change the local
  268. configuration file, with commit and push left to the requested workflow.
  269. ## Resources
  270. - [Documentation](https://docs.coderabbit.ai/cli)
  271. - [Support](https://docs.coderabbit.ai/support)
  272. - [Privacy Policy](https://www.coderabbit.ai/privacy-policy)
  273. - [Terms of Service](https://www.coderabbit.ai/legal/terms-of-service)
  274. - [CodeRabbit Documentation](https://coderabbit.ai/docs)
  275. - [Vercel Skills CLI](https://github.com/vercel-labs/skills)
  276. - [Agent Skills Specification](https://agentskills.io/specification)
  277. - [Gemini CLI Extension Documentation](https://geminicli.com/docs/extensions/reference/)
  278. - [Antigravity CLI Plugin Documentation](https://antigravity.google/docs/cli-plugins)
  279. ## License
  280. MIT

Version history

  • v1.3.011/13/2026

  • v1.3.211/12/2026

  • v1.2.010/19/2026

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.