mindspeed-mm-generative

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill explicitly requires cloning public repositories (e.g., git clone https://gitcode.com/Ascend/MindSpeed-MM.git and https://github.com/NVIDIA/Megatron-LM.git), downloading wheels and model weights from public sources (Ascend PyTorch releases on gitcode and HuggingFace weights such as Wan-AI/Wan2.1-T2V-1.3B-Diffusers), and then loading/executing those artifacts (e.g., model_t2v.json "from_pretrained" paths and feature_extraction.sh → get_wan_feature.py) as mandatory runtime steps, so untrusted third‑party content is fetched and consumed in the workflow and can influence tool execution.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly requires cloning and installing external repositories that will be fetched and executed at runtime (for example: git clone https://gitcode.com/Ascend/MindSpeed-MM.git), so these URLs introduce remote-code-execution risk as required runtime dependencies.