Google puts new limits on AI agents that hunt threats and fix vulnerabilities

Google Cloud is adding new controls to CodeMender, its AI tool for finding and fixing software vulnerabilities, allowing businesses to isolate attack simulations and target code scans more precisely.

The updates let businesses verify vulnerabilities and simulate exploits in an operating system-level sandbox or an isolated virtual machine. They also add human checks for local writes and command execution to help keep testing within defined limits.

Businesses can scan an entire code repository, a specific folder, or only files modified since the last scan, giving them more control over scanning speed and costs. CodeMender remains in preview.

“As adversaries attack at machine speed, Google’s Fairwind Program helps us keep pace with our application security. Our early evaluation of CodeMender shows the potential to reduce vulnerability discovery time by over 95%, while keeping expert validation and remediation under human oversight,” said Ivan Fontarensky, CTO Cybersecurity Services, Thales Group.

New controls for agents and cloud workloads

Google Cloud is extending common security and governance controls across environments where AI agents run, including its Gemini platform, Google Kubernetes Engine (GKE), Cloud Run, and the new Agent Compute Engine.

The controls are designed to help businesses manage agent credentials, protect company data, and detect threats across cloud environments.

Auth Manager, now generally available, centralizes credential management in a Google-managed vault governed by IAM, and handles authorization flows and refreshes user-delegated tokens.

Agent Identity JWT support, available in preview, lets agents authenticate across services including AWS, Azure, on-premises backends, and Model Context Protocol (MCP) servers without relying on static, long-lived API keys.

Google is also expanding identity-based controls for data moving across protected cloud boundaries. VPC Service Controls ingress and egress rules now support IAM roles, generally available, and Workload Identity Federation for GKE, available in preview.

A new Model Armor custom topic detector, also in preview, lets businesses define prompt-based rules to identify whether agent conversations stay within approved topics.

The network security updates include the generally available Cloud NGFW advanced malware sandbox, powered by Palo Alto Networks WildFire. The company is making Cloud NGFW project-level resources generally available and introducing Cloud Armor managed rules in preview, using threat intelligence from Thales Imperva to help protect applications against exploits.

Agent-assisted threat detection and response

Google says its approach to countering autonomous, multi-agent threats combines continuous workload monitoring, threat intelligence, and automated investigations.

Continuous monitoring detects active compromises at the workload layer and feeds telemetry into the Wiz Security Graph. Google Threat Intelligence telemetry is incorporated into defensive models to trigger automated threat hunts that reconstruct process trees and isolate threats in real time, according to the company.

Autonomous agents analyze alerts, run correlation queries across telemetry from multiple vendors, and present mitigation actions for human approval.

Google is also positioning Mandiant Transformation Services as a way to help organizations adapt their security operations, processes, and workforce skills to use these technologies and maintain business continuity during a breach.

Extending security across multicloud environments

Google and Wiz are combining their security tools to help organizations identify and address risks across code, applications, and multicloud infrastructure.

“Together, Wiz and Google Cloud are combining deep autonomous application understanding, connected security context, and threat intelligence to help organizations innovate securely and stay ahead of emerging threats,” Sandra Joyce, VP, Google Threat Intelligence, explained.

Wiz discovers and maps autonomous applications, agents, and infrastructure, analyzing applications’ purpose, capabilities, and permissions to uncover hidden risks. Context from the Wiz Security Graph helps reveal complex attack paths and runtime threats so teams can prioritize remediation.

Now in preview, Wiz AI Static Application Security Testing (SAST) correlates code weaknesses with infrastructure context in the Wiz Security Graph. Teams can then use Green Agent and CodeMender to remediate exploitable attack paths.

A shared data model now connects Wiz Defend and Google Security Operations, allowing security teams to investigate threats in either platform and synchronize information between them. Analysts can also access Wiz Blue Agent’s investigation findings within Google Security Operations, while Wiz Defend now integrates with Google Threat Intelligence.

Google DeepMind and Wiz have also launched Scan for Good, which combines automated tools with human security researchers to help organizations with limited security resources. Google says the initiative used Gemini 4 Argon to help more than 50 public service organizations and critical infrastructure providers identify and address critical security exposures during its first month.