Security flaws found in top open source AI gateway — LiteLLM vulnerabilties could affect millions of users
A patch has been made available, so make sure to apply it before someone racks up a big AI bill.
Two critical vulnerabilities have been discovered in the widely used open-source AI gateway called LiteLLM, posing a significant risk to millions of users. The security flaws, uncovered by Check Point Research, enable threat actors to gain administrative control, steal sensitive data, and execute malicious code remotely. LiteLLM, which boasts over 90 million monthly downloads on GitHub, connects applications to various large language models (LLM) providers, such as OpenAI and Anthropic, through a single interface.
This allows users to route requests, manage API keys, control user access, and monitor usage and costs. The first vulnerability, CVE-2026-35029, is an access control bug that allows low-privileged users to bypass security measures and access administrative functions. This can result in unauthorized file access, credential manipulation, and malware deployment.
The second vulnerability, CVE-2026-59823, is a server-side request forgery bug that enables users to send unauthorized API requests to third-party servers, potentially stealing AI provider authentication keys and causing costly AI account usage. To address these issues, LiteLLM users should upgrade to version 1.83.9 or higher and rotate any potentially exposed credentials.
Additionally, organizations should review access controls and rotate AI provider API keys if LiteLLM has been compromised. It is crucial to incorporate AI gateway security assessments into broader security testing practices.
Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
