Model & ML supply-chain security

Hugging Face Hub malware and pickle scanning: set of built-in scans on the Hugging Face Hub

Hub scanning is a set of built-in scans on the Hugging Face Hub from Hugging Face. It is a commercial product offered as: SaaS. Pricing model: included with platform.

  • Last reviewed

Best for

Teams that download models from the Hugging Face Hub and want to know what the Hub already checks before they add their own scanning

Ownership

NVIDIA entered into a definitive agreement on September 2, 2026 to acquire Hugging Face. Its Form 8-K says closing is expected in the first half of 2027, subject to customary closing conditions including regulatory approvals. Source, checked October 21, 2026: sec.gov (opens in a new tab)

Consider instead if

  • You need to scan model files in your own pipeline before loading them. See ModelScan.

What Hub scanning does

Hugging Face runs security scans on files stored in Hub repositories, so models are checked where they are shared. The docs say every repository file under 2 GB goes through a ClamAV malware scan at each commit. A separate pickle import scan lists the imports in pickled files such as PyTorch weights and highlights suspicious ones. Results from two third-party scanners, Protect AI Guardian and JFrog, also appear on the Hub. The docs say the pickle scan is not 100% foolproof and users remain responsible for checking safety. This is a hosted Hub feature for consumers of Hub models, not a scanner you run in your own pipeline.

This summary is our own wording of the sourced facts on this page.

Key facts

Each value links to the page that states it, with the date we checked it.

Ownership
NVIDIA entered into a definitive agreement on September 2, 2026 to acquire Hugging Face. Its Form 8-K says closing is expected in the first half of 2027, subject to customary closing conditions including regulatory approvals. Source, checked October 21, 2026: sec.gov (opens in a new tab)
Latest release
Not documented in the sources we checked
API reference
Not documented in the sources we checked
Quickstart
Not documented in the sources we checked
Attestations or disclosure
SOC2 Type 2 certified; GDPR compliant Source, checked October 21, 2026: huggingface.co (opens in a new tab)

Capabilities

As documented by the vendor or project. The tag is our label from a controlled vocabulary.

Where Hub scanning fits in our AI security layers

Our analysis

Our editorial model, not a standard. The primary layer follows the category, except for tools whose main documented job is detection and response; other layers rest on a capability the sources document. How the model works.

AI security layers that Hub scanning covers, with the role, how each applies and what it rests on
LayerRoleHow it appliesBased on
L3 Supply chain Model and supply chain Primary Check that models, datasets and AI components from outside are what they claim to be and do not run code or hide backdoors when loaded. Follows its category, Model & ML supply-chain security

OWASP LLM Top 10 mapping

Our analysis

The risk list is OWASP's (Top 10 for LLM Applications, 2025). The mapping is our analysis, and each row rests on a capability the sources document.

OWASP LLM risks mapped to Hub scanning, with the capabilities each mapping rests on
RiskHow it appliesBased on
LLM03 Supply Chain Checks model files in Hub repositories for malware and risky pickle imports before users download them. Runs repository files under 2 GB through a ClamAV malware scan at each commit; Lists the imports in pickled files and highlights suspicious ones on the file page; Scans files pushed to the Hub and warns users when a repository has a file scanned as unsafe; Shows results from third-party scanners from Protect AI (Guardian) and JFrog on the Hub

Alternatives and related tools

More in Model & ML supply-chain security, ranked within the category by Coverage Score.

ModelAudit (Promptfoo)

OpenAI (Promptfoo)

Open sourceMIT
  • CLI
  • Library
  • Self-hosted

Scans model files for malicious code, backdoor indicators and unsafe operations without loading them.

Best for: Teams that want a free scanner covering many model and archive formats, with SARIF output for code scanning in CI

Coverage Score 86 out of 100
  • LLM03 Supply Chain
  • LLM04 Data and Model Poisoning

Reviewed

ModelScan

Palo Alto Networks (Protect AI)

Open sourceApache-2.0
  • CLI
  • Library

Scans saved ML model files for unsafe serialized code before you load them, from a CLI or Python.

Best for: ML and platform teams that want a free scanner to check downloaded or internally produced model files before loading or deploying them

Coverage Score 84 out of 100
  • LLM03 Supply Chain
  • LLM04 Data and Model Poisoning

Reviewed

Open sourceApache-2.0
  • CLI
  • Library

Signs machine learning model files and verifies them later so tampering can be detected.

Best for: Teams that publish or consume models and want a free way to sign model files and check signatures before loading them

Coverage Score 78 out of 100
  • LLM03 Supply Chain
  • LLM04 Data and Model Poisoning

Reviewed

Hub scanning FAQ

What does the Hugging Face Hub scan for?

The docs describe a ClamAV malware scan on repository files under 2 GB at each commit, and a pickle import scan that lists and flags suspicious imports.

Is the Hub's pickle scan a guarantee of safety?

No. The docs say the scan is not 100% foolproof and that users are responsible for checking whether a file is safe.

Can I run this scanner myself?

The docs describe a hosted Hub feature, not a downloadable scanner. For your own pipeline, use a separate scanner such as ModelScan.

Who owns Hugging Face?

NVIDIA announced on September 3, 2026 that it has agreed to acquire Hugging Face. NVIDIA's blog post and Form 8-K do not say the deal has closed, so Hugging Face is still listed as the vendor here.

Key facts you can cite

We have no commercial relationship with Hugging Face.

Sources & last reviewed

Last reviewed . We compile from the vendors' own public pages and repositories and we do not run the products. Source links open in a new tab. Found a mistake? Report a correction.