CISA Adds LiteLLM Vulnerability to Known Exploited Vulnerabilities Catalog

CISA has added CVE-2026-42271, a command injection vulnerability in BerriAI's LiteLLM open-source AI gateway, to its Known Exploited Vulnerabilities catalog after evidence of active exploitation. LiteLLM is widely used to route requests to various LLM providers via OpenAI-compatible interfaces, making it a critical piece of AI infrastructure. The flaw allows an authenticated user with a valid proxy API key to execute arbitrary commands on the LiteLLM host. The risk is amplified when chained with CVE-2026-48710, a separate vulnerability in the Starlet library, potentially enabling unauthenticated remote code execution. The vulnerability affects organizations deploying LiteLLM as a gateway between applications, users, and model providers. BerriAI has released patches, but unpatched systems remain at risk.

Global Impact

This vulnerability highlights a systemic risk in the rapidly expanding AI infrastructure layer, where open-source tools like LiteLLM are often deployed with minimal security oversight. Economically, the exploitation could lead to data breaches, service disruptions, and increased cybersecurity spending for affected enterprises.