Masatoshi Yoshizawa

Name of the Vulnerable Software and Affected Versions: Sockeye versions prior to 2.3.24 Description: Sockeye is an open-source sequence-to-sequence framework for Neural Machine Translation built on PyTorch. It uses YAML to store model and data configurations on disk. The issue arises from unsafe YAML loading in versions below 2.3.24, which can execute arbitrary code embedded in config files. An attacker can add malicious code to the config file of a trained model and attempt to convince users to download and run it. If users run the model, the embedded code will run locally. Recommendations: For versions prior to 2.3.24, update to version 2.3.24 or above to fix the issue. As a temporary workaround, consider avoiding the use of config files from untrusted sources and restricting the execution of models with potentially malicious config files.

**Name of the Vulnerable Software and Affected Versions** Nameko versions prior to 2.14.0 Nameko versions v3.0.0rc0 through v3.0.0rc9 **Description** The issue allows for arbitrary code execution when deserializing the config file. This can be achieved by tricking Nameko into deserializing a malicious YAML config file. For example, a malicious.yaml file can contain code that executes system commands, such as ` import ('os').system('cat /etc/passwd')`. This can lead to the execution of arbitrary system commands. **Recommendations** For versions prior to 2.14.0, update to version 2.14.0 or later to resolve the issue. For versions v3.0.0rc0 through v3.0.0rc9, update to v3.0.0rc10 or later to resolve the issue. As a temporary workaround, consider only using config files that you trust to avoid exploitation.