Anthropic Launches AI Code Review Tool to Address AI-Generated Code Vulnerabilities

AI-powered 'vibe coding' speeds up development but introduces new vulnerabilities, security risks, and unreadable code. Anthropic has launched Code Review, an AI tool to detect issues before code enters the repository, now available in Claude Code.

Integrated with GitHub, it automatically analyzes pull requests, marks potential problems with red (high priority), yellow (potential issues), and purple (related to existing code), focusing on logical errors instead of style. It uses multi-agent parallel analysis, with a master agent summarizing results.

Currently in research preview for Claude team and enterprise users, targeting large companies like Uber and Salesforce. Priced by token, averaging $15-25 per review, it meets the demand from surging AI-generated code volumes.