Anchor0221

**Name of the Vulnerable Software and Affected Versions** LLaMA-Factory version prior to 1.0.0 **Description** LLaMA Factory enables fine-tuning of large language models. A critical issue exists in the `llamafy baichuan2.py` script, which performs insecure deserialization using `torch.load()` on user-supplied `.bin` files from an input directory. An attacker can exploit this behavior by crafting a malicious `.bin` file that executes arbitrary commands during deserialization. **Recommendations** For versions prior to 1.0.0, update to version 1.0.0 to resolve the issue. As a temporary workaround, consider disabling the `llamafy baichuan2.py` script or restricting access to the input directory to minimize the risk of exploitation. Avoid using the `torch.load()` function on untrusted `.bin` files until the issue is resolved.

**Name of the Vulnerable Software and Affected Versions** ms-swift version 3.3.0 **Description** A remote code execution (RCE) vulnerability exists due to unsafe deserialization in `tests/run.py` using `yaml.load()` from the PyYAML library (versions = 5.3.1). An attacker controlling the YAML configuration file passed to the `--run config` parameter can execute arbitrary code during deserialization, potentially leading to full system compromise. The vulnerability is triggered when a malicious YAML file is loaded, allowing the execution of arbitrary Python commands such as `os.system()`. **Recommendations** Upgrade PyYAML to version 5.4 or higher. Use `yaml.safe load()` instead of `yaml.load()` to mitigate the issue.