Claude Code team details tiered code review process
What happened
On October 15, the claude.dev developer blog described how the Claude Code team handles code review: most code reading goes to automated review, human approval is tiered by risk, and staged rollouts plus fast rollback contain errors. Low-risk changes can merge without human approval once automated checks pass, while medium and high-risk changes need one and two human approvers. From June to September, merged PRs more than doubled and the share needing rollback fell by over 40%. Past incidents feed a daily eval, and after three review rounds the team must reconsider the approach.
Written by AI from the coverage · updated 2 hours ago
Coverage
Follow the reports to see the story from different sides.
- claude.dev BlogPickHow the Claude Code team does code review
The Claude Code team hands most code reading to automated review, sets human approval by risk level, and controls errors with staged rollouts and fast rollbacks. From June to September, merged PRs more than doubled while the share needing rollback fell more than 40%. Low-risk changes can merge after automated checks pass with no human approval; medium and high risk need one and two approvers. Past incidents feed a daily eval, and review that reaches three rounds forces the team to reconsider the approach.
Heat over time
Not enough continuous observations to draw a trend yet.