weibo-transcribe

Pass

Audited by Gen Agent Trust Hub on Oct 11, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads AI models from ModelScope (iic/SenseVoiceSmall). This is a standard and expected operation for the skill's primary purpose of transcription.
  • [COMMAND_EXECUTION]: The skill uses yt-dlp via a subprocess (wrapped in a common utility) to download audio from Weibo. This is a functional requirement for the skill.
  • [DYNAMIC_EXECUTION]: The skill initializes an AI model using AutoModel.from_pretrained with trust_remote_code=True in SKILL.md. While this flag can be risky, it is a standard requirement for many community-contributed models on ModelScope and HuggingFace, and the source (iic/SenseVoiceSmall) is well-known.
  • [SAFE]: The project uses local utility modules from the chubby_common package, which appears to be a local shared library used by the author for common tasks like configuration, Markdown generation, and yt-dlp handling. No malicious patterns or data exfiltration were detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 11, 2026, 07:56 AM