hy-world-2-0-3d-world-model
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions do not contain any malicious patterns, prompt injections, or unauthorized access attempts.
- [EXTERNAL_DOWNLOADS]: Downloads source code, model weights, and dependencies from official and well-known sources, including Tencent's GitHub repository, Hugging Face, and the official PyTorch registry. These operations are documented as standard setup procedures for the HY-World 2.0 framework.
- [COMMAND_EXECUTION]: Provides standard shell commands for environment configuration (Conda) and library installation (pip). No suspicious or obfuscated command sequences were found.
- [DATA_EXPOSURE_EXFILTRATION]: Mentions the use of environment variables for Hugging Face tokens as a configuration step, which is a standard security practice for managing access to gated models. No actual credentials or sensitive data are exposed.
Audit Metadata