Google Opens CodeMender AI Security Agent to Expert Testers

Google is pulling back the curtain a bit further on CodeMender, the AI agent it built to hunt down security flaws in code. The company first unveiled the tool back in October, branding it an "AI agent for code security." Now it's expanding access by inviting handpicked groups of experts to test the API directly.

The pitch is ambitious: Google wants CodeMender to "help secure the world's code bases." That's a massive scope, but given the endless parade of vulnerabilities plaguing software everywhere, an AI-powered security agent isn't exactly a hard sell.

By opening the API to outside experts, Google gets real-world feedback before a wider rollout. It's a calculated move — stress-test the tool with people who actually know what they're looking at before unleashing it on the broader developer community.