CodeMender: Google AI Fixes Security Bugs & Patches

Google has launched a preview of CodeMender, an AI agent designed to find security flaws and create fixes for developers. This tool scans code, confirms if vulnerabilities are exploitable, and then generates patches. Google describes CodeMender as a response to attackers already using AI to speed up their work, aiming to provide defenders with similar automation. It can help reduce zero-day risk by moving from passive scanning to automated code remediation. CodeMender is available through the Gemini Enterprise Agent Platform or as part of AI Threat Defense. A separate version is currently limited to a small number of governments and trusted partners. The agent works in three stages. First, it scans code for vulnerabilities across various languages. Second, it builds and runs a proof-of-concept exploit in a sandbox to confirm flaws and reduce false positives. Finally, it writes a patch and checks it before sending it to a developer for manual approval. This innovation could lead to a more continuous, self-healing software development process.