CodeRabbit

Plan

1. SupportsWork planning: Supports

   Work planning, supports. On Team and above, CodeRabbit Plan turns issues and descriptions into Coding Plans based on your codebase. A coding agent implements them. The issue tracker keeps the planning record.

Build

1. Codebase context: Supports

   Codebase context, supports. Reviews draw on a code index, Learnings, linked issues, MCP servers, and linked repositories for cross-repository changes, all in service of the review.

2. Coding: Supports

   Coding, supports. Fix comments applies the review comments you select. Autofix, renamed from Autopilot on Oct 14, 2026, keeps fixing review feedback, CodeRabbit findings, and failing CI checks until the pull request is ready. The old command names work as aliases for a limited time. The Coding Agent implements tasks from findings or plans and bills by agent minutes while in beta. Other coding agents handle feature development.

3. TouchesAgent runtime: Touches

   Agent runtime, touches. The IDE extension and CLI put reviews inside VS Code-compatible editors and coding-agent loops. The agent the developer steers is another product.

4. Tool execution: Supports

   Tool execution, supports. Each review clones the repository into a sandbox and runs linters and SAST tools there, and it can pull context from Jira, Linear, and connected MCP servers.

Validate

1. Debugging: Touches

   Debugging, touches. The CodeRabbit Agent for Slack, billed on demand, can investigate issues from a thread. CodeRabbit is not a debugging tool.

2. Testing: Supports

   Testing, supports. Team and above add unit test generation. Pre-Merge Checks also verify changes. Your team still needs its test suite and test infrastructure.

3. CodeRabbitCoreCode review: Core

   Code review, core. This is CodeRabbit's primary role. It reviews pull requests and local changes line by line, with summaries, walkthroughs, path-based rules, and Learnings from reviewer feedback.

4. Application security: Supports

   Application security, supports. Every paid CodeRabbit plan runs static analysis and checks for secrets and dependency issues during review. Continuous PR security review requires Advanced. Deep Scan bills by usage. Since Oct 23, 2026, Deep Scan findings include a Reproducer Guide, and Runtime Analysis (closed beta, GitHub Cloud) can run it in a CodeRabbit environment. Automatic verification bills Agent minutes, up to US$6 per finding, to the person who last saved the repository setting.

Ship

1. Source control: Supports

   Source control, supports. Triage ranks open pull requests across repositories. Since Sep 28, 2026, Triage rules can merge low-risk pull requests once required checks and reviews pass, and close stale ones. CodeRabbit can request changes or approve, and authors can approve their own pull requests through it until an admin sets reviews.allow_author_approval to false. The CodeRabbit Security check blocks merges only after it is required in GitHub. The Git platform remains the system of record.

2. CI/CD: Touches

   CI/CD, touches. The CLI can run in CI pipelines and CodeRabbit can analyze failing CI checks. Building and deploying stay with the delivery platform.

Operate

1. Observability: Touches

   Observability, touches. No production telemetry reaches CodeRabbit. Its Slack agent can help investigate an incident only with the context and connections you give it.