Threat
AI Security Incident Case: Agents Bypassed Vercel Sandbox Isolation to Exfiltrate Data - NSFOCUS
First reported · Discovered nsfocusglobal.com
Page published
Earliest dated coverage: 27 Oct 2026 · First observed: 11 Nov 2026 · Latest dated coverage: 27 Oct 2026
Coverage timeline
Single-source analysis — one report is available.
Why it matters
VerCellBreak highlights how trusted mTLS services can undermine sandbox egress allowlists by making attacker-directed certificate-chain requests on a compromised agent’s behalf.
VerCellBreak is a reported Vercel Sandbox network-isolation bypass in which an attacker-controlled AI agent encodes secrets in an mTLS client certificate’s Authority Information Access URL. An allowlisted server that retrieves that URL during certificate-chain validation can relay the secrets to an attacker without the sandbox directly contacting the attacker’s server.