LiteLLM supply chain attack hoovers up passwords and secrets
Attackers distributed a trojanized version (1.107.0) of the popular open-source AI gateway through PyPI and Docker registries on Wednesday. The malware was designed to exfiltrate cloud and AI credentials and included a remote access trojan (RAT) enabling arbitrary code execution. PyPI and Docker registries removed the compromised packages within hours, but anyone running the infected version should assume their secrets were stolen. An earlier LiteLLM compromise in March potentially exposed over 2,500 organizations and 434,000 CI/CD pipelines.
Key points
- Malicious LiteLLM v1.107.0 distributed via PyPI and Docker; safe version is v1.106.2
- Malware steals cloud/AI credentials and includes remote access trojan
- Earlier March attack potentially exposed 2,500+ organizations and 434,000 CI/CD pipelines
Summary and key points written by AI from the headline and a short excerpt, so they can be incomplete or wrong. The Register did the reporting; read the original for the full story.
Read the full story at The Register
