Threat
When LLM Gateways Go Wrong: Findings from LiteLLM - Check Point Research
First reported · Discovered checkpoint.com
Page published
Earliest dated coverage: 13 Oct 2026 · First observed: 8 Nov 2026 · Latest dated coverage: 13 Oct 2026
Coverage timeline
Single-source research — one report is available.
Why it matters
LiteLLM gateway compromise can expose model-provider credentials and give low-privileged users control over shared AI infrastructure, making upgrades to the fixed version a priority.
Check Point Research identified vulnerabilities in LiteLLM v1.77.7, tracked as CVE-2026-35029 and CVE-2026-59823, that were resolved in version 1.83.9. Missing authorization on /config/update allowed low-privileged authenticated users to modify runtime configuration, enabling arbitrary file reads, administrative account takeover, and remote code execution through malicious Python modules loaded from cloud storage. A separate SSRF flaw in the Clientside LLM Credentials feature allowed attackers to redirect API calls and exfiltrate provider keys.