2026-11-16 · CyberNews AI · CRITICAL
LiteLLM & Dify Exploited by Autonomous AI Agents
60-WORD SECURITY BRIEF
Autonomous AI agents exploit CVE-2026-42271 and CVE-2026-48710 in LiteLLM and Dify to exfiltrate cloud API keys. The flaws allow unauthenticated remote code execution via malformed request handling, enabling attackers to read and transmit credentials. Patch developers have released Starlette updates; security teams should apply the latest patches and monitor STS pivot activity.