Introduction
At Hugging Face, security is a top priority. We’re committed to continually enhancing our defenses to safeguard our users and their assets. In this blog post, we’ll take a look at our current security landscape as of August 6th, 2024, and break down key security features available on the Hugging Face Hub.
Default Hub Security Features
The following security features are available to all users of the Hugging Face Hub. We highly recommend that you use all of these controls where possible as it will help increase your resiliency against a variety of common attacks, such as phishing, token leaks, credential stuffing, session hijacking, etc.
Fine Grained Token
User Access Tokens are required to access Hugging Face via APIs. In addition to the standard “read” and “write” tokens, Hugging Face supports “fine-grained” tokens which allow you to enforce least privilege by defining permissions on a per-resource basis, ensuring that no other resources can be impacted in the event the token is leaked. Fine-grained tokens offer a plethora of ways to tune your token.
Two Factor Authentication (2FA)
Two factor authentication adds an extra layer of protection to your online accounts by requiring two forms of verification before granting access. 2FA combines something you know (like a password) with something you have (such as a smartphone) to ensure that only authorized users can access sensitive information.
Commit Signing
Commit signing adds an additional layer of security by using GPG to mitigate the issue of impersonating authors. You can learn more about commit signing here: Git Tools: Signing Your Work.
Organizational Access Controls
Organizations on Hugging Face have access to Organizational Access Controls. This allows teams and businesses to define least privilege access to their organization by assigning “read”, “write”, “contributor” or “admin” roles to each of their users.
Automated Security Scanning
Hugging Face implements an automated security scanning pipeline that scans all repos and commits. Currently, there are three major components of the pipeline:
- malware scanning: scans for known malware signatures
- pickle scanning: scans pickle files for malicious executable code
- secret scanning: scans for passwords, tokens and API keys
In the event a malicious file is detected, the scans will place a notice on the repo allowing users to see that they may potentially be interacting with a malicious repository.
Enterprise Hub Security Features
In addition to the security features available to all users, Hugging Face offers advanced security controls for Enterprise users. These additional controls allow enterprises to build a security configuration that is most effective for them.
Single Sign-On (SSO)
Single sign-on (SSO) allows a user to access multiple applications with one set of credentials. Enterprises have widely moved to SSO as it allows their employees to access a variety of corporate software using identities that are managed centrally by their IT team.
Resource Groups
In addition to the base organizational access controls, Enterprises can define and manage groups of repositories as Resource Groups. This allows you to segment your resources by team or purpose, such as “Research”, “Engineering”, “Production” so that the compromise of one segment can not affect others.
Organization Token Management
Enterprise users can now manage which tokens can access their organization and resources. Organization owners can enforce the usage of fine-grained tokens and require administrator approval for each token. Administrators can review and revoke each token that has access to their repositories at any time.
Data Residency
Enterprise users have access to data residency controls, which allow them to define where repositories (models, datasets, spaces) are stored. This allows for regulatory and legal compliance, while also improving download and upload performance by bringing the data closer to your users.
Audit Logs
Enterprise users have access to audit logs that allow organization admins to review changes to repositories, settings and billing. The audit logs contain the username, location, IP, and action taken and can be downloaded as a JSON file which can be used in your own security tooling.
Compliance
Hugging Face is SOC2 Type 2 certified and GDPR compliant. We offer Business Associate Addendums for GDPR data processing agreements to Enterprise Plan users.
Conclusion
At Hugging Face, we’re committed to providing a secure and trustworthy platform for the AI community. With our robust security features, users can focus on building and deploying AI models with confidence. Whether you’re an individual researcher or a large enterprise, our security features are designed to empower you to protect yourself and your assets.
Frequently Asked Questions
Q: What are the default security features available to all users of the Hugging Face Hub?
The default security features available to all users of the Hugging Face Hub include fine-grained tokens, two-factor authentication, commit signing, organizational access controls, and automated security scanning.
Q: What are the Enterprise Hub security features?
The Enterprise Hub security features include single sign-on (SSO), resource groups, organization token management, data residency, and audit logs.
Q: How do I enable fine-grained tokens?
Enabling fine-grained tokens requires you to log in to your Hugging Face account, go to the “Security” tab, and select the “Fine-Grained Tokens” option.
Q: What is two-factor authentication (2FA) and how do I enable it?
Two-factor authentication (2FA) is an additional layer of security that requires you to provide two forms of verification before granting access to your account. To enable 2FA, you can log in to your Hugging Face account, go to the “Security” tab, and select the “Two-Factor Authentication” option.
Q: How do I manage organization tokens?
Organization token management allows Enterprise users to manage which tokens can access their organization and resources. To manage organization tokens, you can log in to your Hugging Face account, go to the “Security” tab, and select the “Organization Tokens” option.
Q: What is data residency and how do I enable it?
Data residency is a feature that allows Enterprise users to define where repositories (models, datasets, spaces) are stored. To enable data residency, you can log in to your Hugging Face account, go to the “Security” tab, and select the “Data Residency” option.
