ServiceNow AI Platform EUVD-2026-67221

The CVSS vector (AV:N/AC:L/AT:N/PR:N/UI:N) indicates network-reachable, unauthenticated exploitation with no user interaction against an affected ServiceNow AI Platform instance. However, ServiceNow explicitly limits this to 'certain circumstances,' meaning exploitation depends on a specific - but undisclosed in the provided data - configuration, feature, or code path being reachable rather than every instance being trivially exploitable; the exact precondition is not specified in the input and should be obtained from KB3152242. Limiting factors: only affected/unpatched versions are at risk, ServiceNow-hosted instances have already been updated by the vendor (reducing exposure to self-hosted and partner deployments), and the attacker must be able to reach the vulnerable endpoint over the network.

This is a genuine top-priority issue rather than an inflated high-CVSS score: the CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N) describes network-reachable, low-complexity, unauthenticated exploitation with high impact across confidentiality, integrity, and availability of both the vulnerable and subsequent systems, yielding the maximum 10.0. The signals are, however, not uniform - no EPSS score, no KEV listing, and no POC were provided, and the vendor's own description qualifies exploitation as possible only 'in certain circumstances,' so the practical attack surface may be narrower than the raw 10.0 implies. Because ServiceNow is a widely deployed enterprise ITSM/workflow platform holding sensitive operational and identity data, any organization running an affected instance should treat patching as urgent even absent confirmed in-the-wild activity; the absence of KEV reflects no observed exploitation, not low severity. Verify the exact 'certain circumstances' precondition with ServiceNow (KB3152242), since that determines whether a given instance is realistically exploitable.

An unauthenticated attacker who can reach an affected ServiceNow AI Platform instance over the network sends a crafted request to a vulnerable endpoint whose input is passed unsanitized into a backend SQL query. Under the qualifying conditions, the injected SQL executes against the instance database, letting the attacker read or alter records beyond their authorization - for example exfiltrating sensitive tables or tampering with instance data. No authentication or user interaction is required per the CVSS vector; no public POC was identified at time of analysis.