CVE-2026-45804 huggingface CVE debrief

Who should care

Users of the Diffusers library, especially those utilizing custom pipeline code or untrusted repositories, should be aware of this vulnerability and take steps to upgrade or mitigate the risk. This includes organizations using the library for machine learning model deployments, data scientists working with diffusion models, and security teams responsible for monitoring and patching vulnerabilities in their software supply chain. Additionally, anyone relying on the Diffusers library for sensitive data processing or critical infrastructure should prioritize upgrading to version 0.38.0 or later to address the bypass of trust_remote_code guard. Security teams should also monitor for and apply future security updates to the Diffusers library to ensure the security of their environments. Furthermore, developers using custom pipeline code should implement additional security measures, such as code reviews and sandboxing, to reduce the risk of exploitation. Lastly, organizations should consider implementing compensating controls, such as network segmentation and access controls, to limit the potential impact of a successful exploit. Asset inventory management is also crucial to identify and prioritize affected systems for remediation. Rollback change windows should be planned to minimize downtime during the remediation process, and source tracking should be enabled to monitor for potential attacks. By taking these steps, organizations can reduce the risk associated with this vulnerability and protect their environments from potential exploitation. The CVE record was published on 2026-07-15T17:16:48.310Z and has not been modified since then, emphasizing the importance of immediate attention to this issue. The NVD entry is currently Analyzed, providing further context for security teams to assess and prioritize this vulnerability. Overall, a comprehensive approach to vulnerability management, including regular updates, monitoring, and compensating controls, is essential to mitigate the risks associated with this vulnerability in the Diffusers library. The issue's severity and potential impact underscore the need for prompt action to secure affected deployments and prevent

Technical summary

The Diffusers library's DiffusionPipeline.from_pretrained flow can bypass the trust_remote_code guard prior to version 0.38.0. This allows custom pipeline code from a Hub repository to execute through the custom pipeline flow without passing custom_pipeline or trust_remote_code=True. The issue arises because the download() function validates model_index.json and custom pipeline code before later loading from a cached folder that can change. This enables a Hub repository with custom .py pipeline code to execute through the custom pipeline flow without proper validation. The issue is fixed in version 0.38.0, where the trust_remote_code guard is enforced properly.

Defensive priority

Organizations using Diffusers library should prioritize upgrading to version 0.38.0 or later to address the bypass of trust_remote_code guard.

Recommended defensive actions

• Upgrade Diffusers library to version 0.38.0 or later
• Implement trust_remote_code=True when using custom pipeline code
• Monitor for and apply future security updates to the Diffusers library
• Review compensating controls for exposed systems while remediation is scheduled and verified
• Check relevant monitoring, detection, and logs for exposed assets that need extra review
• Track exceptions, retest remediated assets, and close the item only after evidence is documented
• Confirm whether affected product deployments exist in managed environments and assign an owner for follow-up

Evidence notes

The CVE-2026-45804 issue arises from the Diffusers library's DiffusionPipeline.from_pretrained flow, which can bypass the trust_remote_code guard prior to version 0.38.0. This allows custom pipeline code from a Hub repository to execute without requiring custom_pipeline or trust_remote_code=True. The issue is addressed in version 0.38.0.

Official resources