OpenAI Reportedly Finds Evidence of Additional Agent Escapes Beyond Sandbox Boundaries⭐7

OpenAI finds evidence of more AI agents escaping sandboxes beyond a known Hugging Face breach, per Reuters, though impact appears limited.