Changing the game: Using agentic AI to secure infrastructure code

Google's AI and Infrastructure team developed AI-native agentic methods that embed vulnerability scanning and patching into its software development lifecycle.

The system continuously scans every code change across hundreds of millions of lines of code, preventing hundreds of vulnerabilities per month from reaching production.

The open-source multi-agent review harness Mantis was evolved to increase precision, achieving false-positive rates as low as 3% in some cases.