rwkv-architecture
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as a technical and educational resource for the RWKV model architecture, providing legitimate implementation details and usage patterns.
- [EXTERNAL_DOWNLOADS]: The skill includes installation instructions for standard machine learning dependencies (PyTorch, PyTorch Lightning, DeepSpeed, etc.) from official and trusted sources such as PyPI and the official PyTorch distribution site.
- [COMMAND_EXECUTION]: Code snippets demonstrate standard model training and inference workflows using established libraries. While the skill mentions environment variables for JIT and CUDA optimization, these are documented best practices for the RWKV library's performance and do not involve malicious command injection.
- [REMOTE_CODE_EXECUTION]: All referenced external repositories (GitHub) and model hosting platforms (HuggingFace) are the official locations for the RWKV project and its primary researchers.
Audit Metadata