CVE-2023-43654
TorchServe Server-Side Request Forgery
Severity Score
Exploit Likelihood
Affected Versions
Public Exploits
2Exploited in Wild
-Decision
Descriptions
TorchServe is a tool for serving and scaling PyTorch models in production. TorchServe default configuration lacks proper input validation, enabling third parties to invoke remote HTTP download requests and write files to the disk. This issue could be taken advantage of to compromise the integrity of the system and sensitive data. This issue is present in versions 0.1.0 to 0.8.1. A user is able to load the model of their choice from any URL that they would like to use. The user of TorchServe is responsible for configuring both the allowed_urls and specifying the model URL to be used. A pull request to warn the user when the default value for allowed_urls is used has been merged in PR #2534. TorchServe release 0.8.2 includes this change. Users are advised to upgrade. There are no known workarounds for this issue.
TorchServe es una herramienta para servir y escalar modelos de PyTorch en producción. La configuración predeterminada de TorchServe carece de una validación de entrada adecuada, lo que permite a terceros invocar solicitudes de descarga HTTP remotas y escribir archivos en el disco. Se podría aprovechar este problema para comprometer la integridad del sistema y los datos confidenciales. Este problema está presente en las versiones 0.1.0 a 0.8.1. Un usuario puede cargar el modelo de su elección desde cualquier URL que desee utilizar. El usuario de TorchServe es responsable de configurar las URL permitidas y especificar la URL modelo que se utilizará. En PR #2534 se fusionó una solicitud de extracción para advertir al usuario cuando se utiliza el valor predeterminado para Allow_urls. La versión 0.8.2 de TorchServe incluye este cambio. Se recomienda a los usuarios que actualicen. No se conocen workarounds para este problema.
The PyTorch model server contains multiple vulnerabilities that can be chained together to permit an unauthenticated remote attacker arbitrary Java code execution. The first vulnerability is that the management interface is bound to all IP addresses and not just the loop back interface as the documentation suggests. The second vulnerability (CVE-2023-43654) allows attackers with access to the management interface to register MAR model files from arbitrary servers. The third vulnerability is that when an MAR file is loaded, it can contain a YAML configuration file that when deserialized by snakeyaml, can lead to loading an arbitrary Java class.
CVSS Scores
SSVC
- Decision:Attend
Timeline
- 2023-09-20 CVE Reserved
- 2023-09-28 CVE Published
- 2024-06-06 First Exploit
- 2024-09-23 CVE Updated
- 2024-11-11 EPSS Updated
- ---------- Exploited in Wild
- ---------- KEV Due Date
CWE
- CWE-918: Server-Side Request Forgery (SSRF)
CAPEC
References (6)
URL | Tag | Source |
---|---|---|
https://github.com/pytorch/serve/pull/2534 | Issue Tracking | |
https://github.com/pytorch/serve/releases/tag/v0.8.2 | Release Notes | |
- |
URL | Date | SRC |
---|
URL | Date | SRC |
---|---|---|
https://github.com/pytorch/serve/security/advisories/GHSA-8fxr-qfr9-p34w | 2023-10-31 |
Affected Vendors, Products, and Versions
Vendor | Product | Version | Other | Status | ||||||
---|---|---|---|---|---|---|---|---|---|---|
Vendor | Product | Version | Other | Status | <-- --> | Vendor | Product | Version | Other | Status |
Pytorch Search vendor "Pytorch" | Torchserve Search vendor "Pytorch" for product "Torchserve" | >= 0.1.0 < 0.8.2 Search vendor "Pytorch" for product "Torchserve" and version " >= 0.1.0 < 0.8.2" | - |
Affected
|