A critical security flaw in ServiceNow’s AI Platform is now being exploited by attackers, raising the risk for thousands of organizations that depend on the software to run IT, HR, customer service, and other core business operations.
Security researchers say the vulnerability allows attackers to execute malicious code without first logging in, making it one of the most serious threats facing ServiceNow customers this year.
Tracked as CVE-2026-6875, the vulnerability is a pre authentication remote code execution vulnerability with a CVSS score of 9.5. It affects the ServiceNow AI Platform and allows an attacker to break out of the platform’s sandbox and execute code on a vulnerable instance. ServiceNow has already released security updates and is urging customers to apply them as quickly as possible.
Attackers Moved Quickly after the Flaw Became Public
The vulnerability was publicly disclosed on July 13, and according to threat intelligence firm Defused, attackers began exploiting it within days of the disclosure.
Researchers said the attacks targeted the same exposed endpoint (/assessment_thanks.do) documented in the public proof of concept. However, the attackers used a different sandbox escape technique to achieve remote code execution instead of relying on the exact method described in the published exploit.
This means security measures designed only to detect the original proof of concept may not identify these attack variants. Later reporting also pointed to additional sandbox escape chains being used, suggesting attackers are already adapting the exploit as more details become public.
Why This Flaw is a Major Concern
Unlike many enterprise software vulnerabilities, this one does not require authentication, meaning an attacker does not need valid credentials before attempting to exploit it.
If the attack succeeds, the attacker can escape the platform’s sandbox and execute arbitrary code. Because ServiceNow is commonly used to manage IT operations, HR workflows, customer service, and other critical business functions, a successful compromise could provide attackers with access to highly sensitive enterprise environments.
ServiceNow is used by many of the world’s largest organizations, making the platform an attractive target whenever critical vulnerabilities emerge. The combination of broad adoption, remote exploitation, and active attacks has made this flaw a top priority for enterprise security teams.
ServiceNow Says Patches Are Available
ServiceNow has released patches for affected customers and said it has also strengthened sandbox restrictions to limit the types of code that can run inside the platform.
The company said its investigation has not found evidence that the reported exploitation affected ServiceNow hosted customer instances, but it still recommends that every customer install the latest security updates without delay.
The rapid exploitation of CVE-2026-6875 is another reminder that attackers closely monitor newly disclosed vulnerabilities and often begin targeting them before many organizations have time to respond.
In this case, attackers are not simply copying a published proof of concept. They are adapting it by using different sandbox escape techniques to reach the same outcome. That makes the campaign harder to detect and increases the urgency for organizations to patch vulnerable systems as quickly as possible.
