Claude Code /code-review Adds Depth Levels and New Review Strategies
In /code-review from Claude Code, analysis depth levels are now available, and the review process has been redesigned for each level.
At the low level, /code-review finds more issues than other code review tools, at a significantly lower token cost. The high level delivers noticeably higher recall when deeper analysis is needed.
You choose the trade-off.
Previously, /code-review used a single fixed prompt regardless of the selected depth level. Now, a separate review strategy is applied for each level. The command automatically uses the depth level set for the current session, or it can be specified explicitly, for example: /code-review high.
At the low level, a single pass over the diff is performed. This is fast and cheap enough to run a check before every push.
At the medium level, the changed code is analyzed in context, then multiple passes are run to look for issues from different angles, after which each found issue is verified before it makes it into the final report.
At the high level, search and verification agents are launched as sub-agents with a clean context, so they are not tied to the reasoning of the agent that just wrote the code. At the x-high level, the review goes even deeper and covers the possible impact of changes on code beyond the diff itself.
There is one more level above high: /code-review ultra. It launches an entire group of reviewer agents and independently reproduces each found issue. Coverage of critical issues remains the same as at the high level, but there are significantly fewer false positives.
At Anthropic, they run this level for every PR.