Google Cloud adds sandbox controls to CodeMender and its AI agents

Feeds.4Sysops •IT News • November 5, 2026

CodeMender is Google Cloud’s AI agent for finding and fixing software vulnerabilities. New controls let organizations isolate its exploit testing and choose how much code it scans, while broader Google Cloud updates add security guardrails for agents running across the company’s platforms. Source

The full story

This article is one source in a clustered incident — the cluster page carries the summary, timeline and every other outlet covering it.

Google Cloud Enhances CodeMender with New Security Controls Google Cloud has introduced new security controls for CodeMender, its AI tool designed to identify and fix software vulnerabilities. The updates allow businesses to isolate attack simulations and conduct targeted code scans, enhancing control over scanning speed and costs. These controls include the ability to verify vulnerabilities within an operating system-level sandbox or an isolated virtual machine, and they incorporate human oversight for local writes and command execution. Additionally, Google Cloud is extending security and governance measures across its platforms, including the Gemini platform and Google Kubernetes Engine. The enhancements aim to bolster security for businesses utilizing AI agents in cloud environments, with features like centralized credential management and identity-based controls for data movement. CodeMender remains in preview, with early evaluations indicating a potential reduction in vulnerability discovery time by over 95%. 4 sources 31

Google Cloud adds sandbox controls to CodeMender and its AI agents

The full story

Extracted Entities

Platforms (1)

Tools (1)