Critical ServiceNow AI Platform Flaw Exploited for Unauthenticated Code Execution

The window between public disclosure and active exploitation has closed abruptly for enterprise organizations relying on ServiceNow. Threat intelligence reports now confirm that a critical vulnerability within the ServiceNow AI Platform is being weaponized by attackers in the wild, marking a significant escalation in risk for IT service management environments globally. This development transforms a theoretical risk into an immediate operational crisis for security teams tasked with defending these vital infrastructure platforms.

At the center of this emergency is CVE-2026-6875, a security weakness rated with a CVSS score of 9.5, placing it firmly in the critical severity category. According to intelligence shared by Defused Cyber, this vulnerability is classified as a sandbox escape. In technical terms, a sandbox escape allows an attacker to break out of the isolated, restricted environment designed to contain untrusted code or processes, thereby interacting with the broader system. What makes this flaw particularly dangerous is that it permits unauthenticated code execution. This means a malicious actor does not need valid user credentials, API keys, or any prior level of trust within the network to execute arbitrary commands. The barrier to entry for this attack is effectively non-existent, while the potential for damage is extensive.

The implications for security teams are profound and multifaceted. ServiceNow is not merely a ticketing system; it is the central nervous system for many large enterprises, handling IT workflows, human resources data, and customer service operations. A successful sandbox escape in this context could allow an attacker to pivot from the application layer to the underlying host infrastructure. Because the vulnerability is unauthenticated, it bypasses standard perimeter defenses like multi-factor authentication. Security leaders must prioritize the identification of all ServiceNow AI Platform instances within their environment immediately. The concern is not just data exfiltration, but the potential for attackers to establish persistence within the enterprise network, leveraging ServiceNow’s extensive integrations to move laterally to other critical systems.

For defenders, the time to remediate is now. Organizations must review their patch management strategies to ensure they are covering the AI components of their ServiceNow instances, which are sometimes treated separately from core platform updates. Beyond patching, security operations centers should hunt for indicators of compromise, specifically looking for unusual process executions or file modifications within the application logs. Given the high value of the target, it is highly likely that automated exploitation tools are already scanning the internet for vulnerable instances. Consequently, isolating vulnerable systems until patches can be applied may be a necessary temporary measure to prevent breach.

The active exploitation of CVE-2026-6875 serves as a stark reminder of the speed at which threat actors move once a high-value vulnerability is disclosed, particularly within platforms that utilize emerging technologies like AI. The ability to execute code without authentication removes a primary layer of defense, forcing organizations to rely heavily on rapid patching and deep visibility into their network traffic. Security leaders must treat this not just as a routine software update, but as an active incident response scenario. Mitigating this threat requires immediate action to patch systems,