Claude Code Review is Anthropic's managed service that reviews pull requests in your GitHub repositories, using several agents to look for logic errors, security vulnerabilities and broken edge cases. As of October 2026 it's only for Team and Enterprise plans, an Owner of the Claude organization switches it on, and it never approves or blocks a merge.
The catch: the managed reviewer isn't open to everyone
Most people reading about Code Review assume they can switch it on from their own account. You can't unless three things line up. You need a Team or Enterprise plan, your code needs to live on GitHub, and an Owner of the Claude organization has to turn it on. GitHub Enterprise Server and Enterprise Cloud with data residency both count as GitHub.
Some organizations can't use it even on the right plan. As of October 2026 it isn't available to organizations with Zero Data Retention enabled or with the HIPAA configuration applied, it isn't covered by Anthropic's BAA, and it isn't yet available with customer-managed encryption keys (CMEK).
If you're on Pro or Max, or your repositories are on GitLab, you haven't missed out. Claude can still review your code, you just get there a different way. The table further down shows every route.
How a managed review runs, stage by stage
Every review runs on Anthropic's infrastructure in four stages.
- Find. Several agents examine the diff and the code around it in parallel.
- Verify. Findings are checked against the code to filter out false positives.
- Rank. Duplicates are removed and what's left is ordered by severity.
- Post. Results land on the pull request.
You see the findings as inline comments ranked by severity, a summary, and a check run named Claude Code Review. That check run never concludes as a failure. It won't stop a merge, so the decision stays with your people.
Access is read-only. A review doesn't push commits, build your code or run it. It also reads REVIEW.md from the base or default branch, never from the pull request's own changes, so a pull request can't rewrite the rules for its own review.
Review levels and when reviews run
An Owner picks one of three levels for the whole organization, and can give any single repository its own.
- Quick: lighter, the fastest and the lowest cost.
- Standard: the default, balanced between speed and depth.
- Deep: the most thorough, takes longest and costs the most.
There's no model choice in the admin settings. The level is the only dial you get.
Each repository also has a setting for when reviews run: once when a pull request opens, after every push, or only on request. Anyone with write access can type @claude review in a comment to start one. Automatic runs skip drafts and forks, so if you're waiting for a review on a draft pull request, you'll need to ask for it.
Every way to get Claude to review your code
Here's how the managed service compares with the routes that work outside it, as of October 2026.
| Route | What starts it | Where it runs | Who can use it |
|---|---|---|---|
| Code Review (managed) | Pull request activity or @claude review | Anthropic infrastructure | Team and Enterprise, GitHub repos, Owner turns it on |
| /code-review | You run it on your branch | Your Claude Code session | Anyone, any plan |
| Ultrareview | You start it from Claude Code | Cloud sandbox | Needs a claude.ai sign-in |
| Claude in your own CI | Pipeline events, every PR or MR | Your CI runners | API key or cloud credentials, GitHub Actions or GitLab CI/CD |
| Your own subagent or skill | You or Claude invoke it | Wherever Claude Code runs | Anyone who writes one |
For a solo developer, /code-review is the obvious start. You run it in a session on your own branch before you open the pull request, and it works on any plan. If you want every pull request or GitLab merge request checked automatically, running Claude in your own CI does that with an API key. And if your team has its own standards, you can write your own reviewer as a subagent or skill.
If you work in a browser rather than a terminal, my guide to running Claude Code in the browser shows how that side works.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Tuning what Code Review flags
Out of the box, Code Review focuses on correctness: bugs that would break production. It doesn't spend its time on formatting preferences or missing test coverage. That's deliberate, and it keeps the comments worth reading.
If you want it to look for something else, you tune it with REVIEW.md and CLAUDE.md files in the repository. Remember the rule from earlier: REVIEW.md is read from the base or default branch, so a change to it only takes effect once it's merged.
One warning from the docs is worth pinning to the wall. A review with no findings doesn't mean the code is free of errors or secure. Treat a clean review as one signal, not a sign-off.
Once you've sorted this, Claude Code Scheduled Tasks is worth a look.
My verdict on Claude Code Review
I'd switch on the managed service if you're on Team or Enterprise, your code is on GitHub and you want every pull request read before a human looks at it, because the verify step keeps the noise down. I'd start on Standard and only move a repository to Deep if it's handling something sensitive. I'd skip waiting for it if you're on Pro or Max: run /code-review in a session before you push, and you'll have a review today without waiting on anyone. Whatever route you pick, it advises and your team decides. This guide to what each Claude Code plan includes will help if you're choosing between plans.
Frequently asked questions
Can Claude Code Review block a merge?
No. A review never approves or blocks a pull request, and the Claude Code Review check run doesn't conclude as a failure.
Which plans include Claude Code Review?
As of October 2026 the managed Code Review service is for Team and Enterprise plans, on repositories hosted on GitHub, and an Owner of the Claude organization turns it on.
Does Claude Code Review work with GitLab?
The managed service is GitHub only. You can run Claude in GitLab CI/CD with an API key, or use /code-review in a Claude Code session.
How do I start a Claude review on a pull request?
Anyone with write access can comment @claude review on the pull request. Automatic runs skip drafts and forks.
Can I choose the model Code Review uses?
No. The admin settings offer no model choice. You pick a review level: Quick, Standard or Deep.
