claude.com
|
ksl
|
|
Anthropic launched a multi-agent code review system for Claude Code that automatically dispatches parallel agents to analyze pull requests on GitHub. Each review identifies bugs, filters out false positives through verification, and ranks findings by severity – averaging around 20 minutes per PR and costing between $15 and $25 depending on complexity. For large PRs over a thousand lines, the system flagged issues 84 percent of the time, averaging 7.5 findings per review. Available in research preview for Team and Enterprise plans, with per-repo controls and monthly spending caps. The move fits a broader pattern where AI coding tools are expanding from generation into review and QA, a shift that puts pressure on standalone code analysis vendors like Codacy and SonarQube.
