Return to Threats

AI supply-chain risk report described a malicious LiteLLM package on PyPI

KodeKloud 2026-07-21 AI supply chain Critical
Run AI Security Assessment Talk to AI CISO

Executive Summary

KodeKloud reports a malicious LiteLLM release on PyPI in March 2026 put roughly 500,000 credentials at risk, including Meta, OpenAI and Anthropic provider keys.

What Happened

KodeKloud reported that attackers slipped a malicious LiteLLM release onto PyPI in March 2026. The report estimated that the poisoned dependency put roughly 500,000 credentials at risk, including provider keys associated with Meta, OpenAI, and Anthropic.

Affected Products

  • LiteLLM
  • PyPI

Vulnerability / Attack Technique

Poisoned AI gateway dependency

Technical Mechanism

A malicious release harvested credentials on install.

AI Security Category

LLM03:2025 Supply Chain

Part of the AI Supply Chain and AI SBOM topic.

Relevant Frameworks

LLM03:2025 Supply ChainMITRE ATLAS: Credential Access

RealGround Analysis

AI gateways hold keys to many providers at once. RealGround analysis: rotate all LLM provider keys after such incidents.

Recommended Controls

  • Pin LiteLLM versions
  • Rotate LLM provider keys

Related RealGround Service

AI Agent SBOM and LLM Mapping · or start with the free AI Agent Security Scanner.

Sources

KodeKloud: https://kodekloud.com/blog/ai-supply-chain-security-cicd/

Summary and analysis written by RealGround from the source above; last updated 2026-10-18.

Talk to AI CISO