Critical Hugging Face Transformers flaw ran attacker code on a routine model load
A critical zero-day in Hugging Face Transformers just bypassed the industry's standard safety guardrail. Here's what it means for your AI supply chain.
Why it matters
A remote code execution vulnerability in Hugging Face's widely-used Transformers library defeated its own recommended security controls, exposing organizations to arbitrary code execution through model loading—a routine operation in AI development workflows. This represents a critical supply-chain risk for any team downloading models from open repositories.
The key facts
7 to knowCVE-2026-4372
Vulnerability: remote code execution via model loading
Defeated trust_remote_code=False safeguard
Disclosed by Pluto Security Inc.
Affects standard model-loading commands
Impacts organizations following recommended Hugging Face security guidance
Supply-chain security risk for open-source AI workflows
Go to the source
SiliconAnglesiliconangle.com
Publisher excerpt: Pluto Security Inc. today disclosed a critical remote code execution vulnerability in Hugging Face Inc.’s Transformers library that allowed attacker-controlled artificial intelligence models to run arbitrary code on a victim’s machine. The flaw fired through a standard model-loading command, even…