Google releases CodeMender AI agent to hunt and fix code flaws

Google has launched CodeMender, a managed AI code security agent, as part of its Gemini Enterprise Agent Platform. Developed by Google DeepMind, it transitions from a research project to an enterprise tool that finds and fixes software vulnerabilities. CodeMender uses advanced reasoning models for deep analysis and verification of code. It verifies the exploitability of vulnerabilities through PoC exploits and integrates directly into developer environments.

Currently in public preview, it supports multiple programming languages and frameworks, while future updates will enhance its governance and integration capabilities.