Security · 1h ago
Critical ServiceNow AI Sandbox Escape Under Active Exploitation
CVE-2026-6875 is a critical sandbox escape in the ServiceNow AI Platform with a CVSS score of 9.5, allowing unauthenticated remote code execution. Exploitation has been observed in the wild since July 21, 2026, targeting the pre-auth endpoint '/assessment_thanks.do'. Patches have been available since June; unpatched self-hosted instances are at immediate risk of full compromise.
Meridian48 take
The pre-authentication nature of this exploit makes it especially dangerous, as there are no login attempts to flag—organizations should prioritize patching over waiting for change windows.
Read the full reporting
CVE-2026-6875: ServiceNow AI Platform Sandbox Escape Under Active Exploitation →
DEV Community
servicenowsandbox-escape