What happened
ServiceNow patched a critical pre-auth RCE/sandbox-escape flaw in its AI Platform on 2026-07-14. Threat intelligence firm Defused reported on 2026-07-18 that it observed in-the-wild exploitation of the vulnerability, with attackers changing tactics from the original Searchlight Cyber proof-of-concept in response to vendor patches and defenses.
Why it matters
The exploitation targets the sandbox isolation layer meant to safely contain AI-driven code execution within an enterprise AI platform — confirming that AI execution sandboxes are now a primary target for real-world attackers, and that sandbox-escape techniques against AI platforms are being actively weaponized rather than remaining theoretical.
Attack vector
A pre-authentication sandbox-escape vulnerability in the ServiceNow AI Platform allows an unauthenticated remote attacker to break out of the platform's execution sandbox and achieve arbitrary remote code execution under certain circumstances.
Affected systems
ServiceNow AI Platform (self-hosted instances; patched hosted instances)
Mitigation
Apply the ServiceNow security update released 2026-07-14 to all self-hosted instances immediately; hosted instances were already patched by the vendor.