CVE-2024-35199
Published: 19 July 2024
Summary
CVE-2024-35199 is a high-severity Exposure of Resource to Wrong Sphere (CWE-668) vulnerability in Pytorch Torchserve. Its CVSS base score is 8.2 (High).
Operationally, exploitation aligns with the MITRE ATT&CK technique Exploit Public-Facing Application (T1190); ranked at the 21.4th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog.
This vulnerability is AI-related — categorised as Deep Learning Frameworks; in the Supply Chain and Deployment risk domain.
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2024-2363
Vulnerability details
TorchServe is a flexible and easy-to-use tool for serving and scaling PyTorch models in production. In affected versions the two gRPC ports 7070 and 7071, are not bound to [localhost](http://localhost/) by default, so when TorchServe is launched, these two interfaces…
more
are bound to all interfaces. Customers using PyTorch inference Deep Learning Containers (DLC) through Amazon SageMaker and EKS are not affected. This issue in TorchServe has been fixed in PR #3083. TorchServe release 0.11.0 includes the fix to address this vulnerability. Users are advised to upgrade. There are no known workarounds for this vulnerability.
- CWE(s)
AI Security AnalysisAI
- AI Category
- Deep Learning Frameworks
- Risk Domain
- Supply Chain and Deployment
- OWASP Top 10 for LLMs 2025
- None mapped
- Classification Reason
- TorchServe is a production serving tool specifically designed for PyTorch models, which is a deep learning framework. The vulnerability affects the deployment of PyTorch inference servers, directly tied to deep learning model serving.
Related Threats
MITRE ATT&CK Enterprise TechniquesAI
Why these techniques?
Exposes unauthenticated gRPC ports (7070/7071) to all network interfaces by default, enabling remote exploitation of the public-facing TorchServe application (T1190) and endpoint denial-of-service attacks such as service exhaustion floods (T1499).
Affected Assets
Mitigating Controls
Likely Mitigating Controls AI
Per-CVE control mapping for this CVE has not run yet; the list below is derived from the weakness types (CWEs) cited in the NVD entry.
Controls whether organization resources are exposed to external system spheres by permitting or prohibiting their use.
The control ensures information is not released into a security sphere where the recipient lacks matching access authorizations.
The control ensures information resources are not exposed to the incorrect (public) sphere through review and authorization.
Protects against data mining that would expose resources to unauthorized spheres by enforcing detection and controls.
Restricts information flows to ensure resources are not exposed to incorrect or unauthorized spheres.
Controlling internal connections prevents exposure of resources to unintended internal spheres.
Knowing exact processing and storage locations helps avoid exposure of resources to incorrect spheres.
The control prevents exposure of the media resource to the wrong security sphere.