Critical LMCache Flaw Lets Unauthenticated Attackers Run Code Remotely

Summary: A critical flaw in LMCache, open-source software that accelerates LLM servers such as vLLM, lets an unauthenticated attacker run code on the cache server by sending a single network message to its ZeroMQ socket in multiprocess mode. JFrog disclosed the bug on October 7 and rated it 9.8 out of 10; it affects versions 0.3.9 through 0.5.5 and was fixed in 0.5.6, released October 15.
What you should know: The server unpacks one message type with pickle before checking its type, so any host that can open a connection runs code as the LMCache process - which ran as root on the project's official container images at disclosure. LMCache's own example Kubernetes deployment listens on every network interface, and JFrog's advisory offers no way to tell whether a server has already been attacked.

Read the original at The Hacker News ↗

Topics: Vulnerabilities & Patching
Mentions: CVE-2026-105192
Related: ⚡ Weekly Recap: Helpdesk 0-Days, Rogue Certs, Shiny Hunters Arrest, AI-Fooling Malware and More Cyber Risks ·
Attackers Exploit UnrealIRCd Flaw That Enables Server-Level Command Injection Cyber Risks · Nov 14, 2026
U.S. CISA adds Check Point and UnrealIRCd flaws to its Known Exploited Vulnerabilities catalog Cyber Risks · Nov 14, 2026