ServiceNow has patched three new CVSS 10.0 vulnerabilities in its AI Platform — the enterprise workflow system used by over 100,000 enterprise AI applications at 85% of Fortune 500 companies. The three flaws (CVE-2026-18885, CVE-2026-18886, and CVE-2026-74820) allow unauthenticated attackers to execute arbitrary code, escalate privileges, and run SQL injection attacks against targeted instances. A fourth vulnerability, CVE-2026-6876 (high severity), enables attackers with basic privileges to achieve remote code execution via a sandbox escape. All four require no user interaction and are rated low-complexity to exploit.
ServiceNow says it is not aware of active exploitation against customer instances and has already patched its cloud-hosted platform. Organizations running self-hosted (on-premises) deployments need to apply patches manually. ServiceNow has a history of serious security flaws being weaponized quickly: in 2024, attackers chained three separate ServiceNow flaws (CVE-2024-4879, CVE-2024-5178, and CVE-2024-5217) to breach private firms and government agencies worldwide, and an earlier 2026 flaw (CVE-2026-6875) was reported as actively exploited in the wild.
How to check if you’re affected
Affected versions include the Xanadu, Yokohama, Zurich, and Australia release families of the ServiceNow AI Platform. Check your instance’s current patch level against ServiceNow’s advisory: patched versions include Xanadu Patch 11 Hot Fix 7a, Yokohama Patch 12 Hot Fix 3b (and later), Zurich Patch 11 and Patch 12, and Australia Patch 4 and Patch 5. If you are on an older release, apply the relevant hot fix as a priority — CVE-2026-18885, CVE-2026-18886, and CVE-2026-74820 are all exploitable by unauthenticated actors.