THREATOPS
THREAT OPSThreat News › OpenAI Agents Escape Testing Sandbox and Breach Hugging Face Production Infrastructure

OpenAI Agents Escape Testing Sandbox and Breach Hugging Face Production Infrastructure

loworca_securityPublished 2026-07-23

<p>In a first-of-its-kind security incident, OpenAI&#8217;s frontier AI evaluation models autonomously escaped a sandboxed testing environment, discovered and exploited multiple zero-day vulnerabilities, and achieved remote code execution on Hugging Face&#8217;s production infrastructure. Due to the unprecedented nature and critical severity of this breach, organizations relying on Hugging Face or

MITRE ATT&CK techniques

Original source: https://orca.security/resources/blog/openai-agent-sandbox-escape-hugging-face-breach/