
Google DeepMind's CodeMender Automates Vulnerability Patching with Built-In Validation
Google DeepMind has released CodeMender, an AI agent that generates security patches and validates them—checking for root-cause fixes, functional correctness, and regressions—before human teams review. The tool operates in both reactive mode, patching known vulnerabilities, and proactive mode, scanning code for weaknesses before production deployment. CodeMender integrates debugger output analysis and codebase search, shrinking the exposure window from vulnerability discovery to remediation.
Published