Critical flaws in widely used LiteLLM gateway allowed full system takeover
Researchers at Check Point found two high-severity vulnerabilities, tracked as CVE-2026-35029 and CVE-2026-59823, in LiteLLM, an open-source tool with almost 100,000 downloads a month that acts as a gateway between applications and AI model providers like OpenAI, Google and Anthropic. The first flaw let any user gain administrative access through a missing authorization check, opening the door to full system takeover and remote code execution, while the second exploited a server-side request forgery technique to redirect the gateway's outbound requests and expose AI providers' API keys. Because such gateways centralize credentials for every AI provider a company uses alongside user role management, an attacker with even a low-privilege account could potentially gain file access, steal login credentials and execute code remotely. A fix shipped in version 1.83.9, and experts recommend updating immediately, auditing who has gateway access, and rotating API keys if unauthorized access is suspected. For a business connecting internal applications to multiple AI providers through such a tool, the case is a reminder that AI infrastructure needs the same security scrutiny as any other critical system.
Sources
This summary was written automatically with AI from the public sources listed above and checked by an independent AI reviewer before publication. For full details, read the original source.