Claude Code Review leaves preview with per-repo review depth

Anthropic has taken Code Review, the code review service in Claude Code, out of research preview and on October 15, 2026 made it generally available on Claude Team and Enterprise plans. The developer account @ClaudeDevs shared the news as well. Code Review has Claude read every pull request opened in a GitHub repository, and the issues it finds land as comments under the lines concerned.

How a review runs

According to the Claude Code documentation, a review runs on Anthropic's infrastructure in four steps. Several agents read the change and the code around it in parallel, looking for logic errors, security vulnerabilities or missed edge cases. A verification step then checks the candidate findings against the code again and drops false positives. The remaining findings are deduplicated, ranked by severity and posted to the pull request as inline comments, a summary and a check run named Claude Code Review. The review does not build or run the code, reads the repository with read-only access and does not push commits. It neither approves nor blocks the pull request, so the merge decision stays with the team. By default it focuses on bugs that would break production rather than formatting preferences, and teams can tune what gets flagged with REVIEW.md and CLAUDE.md files in the repository.

Quick, Standard and Deep

The most visible addition at general availability is review levels. An organization Owner picks one level for the whole organization and can give any repository its own. Quick is a lighter, fastest pass, Standard is the balanced default, and Deep is the most thorough and longest review. Anthropic suggests Quick for routine codebases and Deep for critical ones; the deeper the level, the higher the cost per review. The company says reviews are noticeably faster than during the preview and that findings state the problem, what triggers it and the fix in a few lines. When reviews run is also set per repository: once when a pull request opens, on every push, or only on request. Anyone with write access can also start a review by commenting @claude review on a pull request.

Setup and credits

To turn the service on, an Owner enables Code Review in Claude Code settings, installs the Claude GitHub App and selects repositories; developers do not need to configure anything. GitHub Enterprise Server and GitHub Enterprise Cloud with data residency are supported. According to the announcement, eligible Claude Enterprise organizations can claim at least $10,000 in Code Review credits until January 15, 2027; the credits expire 30 days after they are claimed, and eligibility terms are on Anthropic's support site. Billing and the monthly spend limit are covered in the documentation. For those who cannot use the service, the /code-review command in a Claude Code session runs a local review on any plan.

Code Review works only with repositories hosted on GitHub. It is not yet available to organizations using Zero Data Retention, the HIPAA configuration or customer-managed encryption keys. Anthropic's own documentation also notes that AI can miss bugs and vulnerabilities, and that a review with no findings does not mean the code is error-free or secure. The speed and quality claims come from the company; teams would do well to try it first on a non-critical repository, compare its findings with their own human reviews and watch for comparative results from independent researchers. Disclosure: GlobalFeed also runs on Claude.