Code review
Run /code-review on what changed, ask for inputs that break the code, rank the findings, and give Claude your team's checklist.
-
How Do You Ask Claude What Would Break Your Code? Ask for failing inputs, not an opinion. Level 2: Context
- 1Fresh reviewerhasn't seen your reasoning
- 2Concrete failing inputsranked by damage
- 3You verifydrop what callers already prevent
- 4Failing tests, then fixesfix only what fails
-
How Do You Rank Claude's Code Review Findings? Ask for a fixed scale, and a concrete failure for anything at the top of it. Level 2: Context
- 1Importantfix before merging
- 2Nitminor, doesn't block
- 3Pre-existingreal, but not from this change
-
How Do You Give Claude a Code Review Checklist? Write down the checks your team actually cares about, and put them where the review will read them. Level 2: Context
- 1In the promptthis one review
- 2CLAUDE.md/code-review and the service
- 3REVIEW.mdthe Code Review service only
-
How Do You Make Claude Review Only What Changed? In Claude Code, run /code-review. Level 2: Context
- 1/code-reviewbranch plus uncommitted
- 2/code-review 1234a pull request
- 3/code-review main...my-featureone branch against another
- 4/code-review HEAD~15..HEADthe last 15 commits
- 5/code-review src/autha file or folder
Five ways to get a review
Anthropic's docs list five options. The /code-review command runs in your own Claude Code session on any plan and counts toward your normal usage. The Code Review service is for Team and Enterprise organizations on GitHub: an Owner turns it on, and it posts inline comments on pull requests, billed through usage credits. /code-review ultra runs a deeper multi-agent review in a cloud sandbox and is a research preview. Claude in your own CI runs through GitHub Actions or GitLab CI/CD. Or you can write your own reviewer as a subagent or skill. Most people should start with /code-review, since it needs no setup. There's also /security-review for risks such as injection and data exposure.
What a review can't tell you
A review is one signal. The Code Review service reads the diff and surrounding code without building or running it, never approves or blocks a pull request, and its docs say a review with no findings doesn't mean your code is error-free or secure. So keep your tests and CI, and keep the decisions that need context, such as whether a change belongs in the product at all. Anthropic also warns that a reviewer asked to find problems will always find something, and that fixing every finding adds needless layers. Read each finding before you act on it, and fix the ones you can confirm.
Can Claude review pull requests automatically?
Yes, through Anthropic's Code Review service, on Team and Enterprise plans with GitHub repositories. Reviews can run once when a pull request opens, after every push, or only when someone comments @claude review. You can also run Claude in your own CI with GitHub Actions or GitLab.
How much does Claude Code Review cost?
The service has no flat per-review price. It bills on token usage through usage credits, and cost depends on the review level, the size of the change and how often reviews run. The local /code-review command just counts toward your normal usage.
What is the difference between /code-review and Code Review?
/code-review is a command you run in your own Claude Code session, on any plan, and it reports findings in your conversation. Code Review is a managed service for Team and Enterprise that reviews GitHub pull requests and posts inline comments.
What is REVIEW.md?
REVIEW.md is a file at your repository root holding instructions for Anthropic's Code Review service, such as what to flag, at what severity, and what to skip. The local /code-review command doesn't read it, so rules you want in both places belong in CLAUDE.md.
Does Claude's code review replace a human reviewer?
No. It can catch bugs a person might miss, but Anthropic's service never approves or blocks a pull request, and it may miss issues. A person still decides whether to merge and whether the design is right.