media-lipsync

Pass

Audited by Gen Agent Trust Hub on Jun 27, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references and provides setup instructions for AI models hosted on well-known platforms like GitHub and HuggingFace. These resources originate from established technology organizations, including ByteDance and Kuaishou, and are used for their intended purpose of providing model weights and inference code.
  • [COMMAND_EXECUTION]: The Python script scripts/lipsync.py utilizes subprocess.run to call model inference scripts. This implementation correctly uses list-based command construction, which avoids the security risks associated with shell interpretation and ensures safe handling of user-provided file paths.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 27, 2026, 09:54 AM
Security Audit — agent-trust-hub — media-lipsync