LiteLLM

The LiteLLM 1.107.0 release was compromised by attackers using stolen developer credentials to inject a credential stealer and a backdoor into PyPI wheels and Docker images. The malicious code targets sensitive data including cloud credentials, SSH keys, and AI provider tokens while maintaining persistent remote access.

The author details a transition from managing LLMs as application-specific code to implementing a robust platform engineering layer. This architectural shift focuses on controlling hallucinations, managing token costs, and ensuring secure tool orchestration.

This report details the rising trend of opportunistic scanning targeting AI-related services and infrastructure. It highlights specific threats to unauthenticated Ollama instances and the use of coordinated sweeps to harvest AI API keys from configuration files.