LiteLLM Open-Source Project Compromised for Two Hours in Supply Chain Attack

A brief supply-chain compromise hit the open-source LiteLLM project, with attackers gaining unauthorized access for roughly a two-hour window on October 21 before the project's maintainers and security firm SafeDep contained it.

Key points:

• LiteLLM is a widely-used open-source tool that helps developers work with multiple AI model providers through a unified interface.

• The two-hour window means any organization that pulled updates from LiteLLM during that period may have received compromised code.

• The rapid containment—two hours from detection to resolution—reflects the open-source security community's improving response capabilities.

This incident joins several others this week—Cairn, the PHP vulnerability, the Railway breach—in illustrating how much AI-adjacent infrastructure is now under active attack. Supply chain compromises are particularly concerning because they can affect many downstream users through a single point of failure.

Organizations using LiteLLM should verify their installations against known-good versions and rotate any credentials that may have been exposed during the compromise window.

Why It Matters: Supply chain attacks on AI infrastructure tools can cascade to thousands of downstream applications—making even brief compromises potentially serious for the broader ecosystem.