video-thumbnail-extractor

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS] (SAFE): The skill requires standard Python packages (moviepy, pillow, numpy) which are widely used and trusted in the data science and media processing communities.
  • [DATA_EXFILTRATION] (SAFE): Analysis of video_thumbnail_extractor.py shows only local file system operations (reading input videos and writing output images). There are no network requests or attempts to access sensitive system directories.
  • [COMMAND_EXECUTION] (SAFE): The script uses library-level abstractions for video processing. It does not invoke system shells or execute arbitrary commands provided by users.
  • [INDIRECT_PROMPT_INJECTION] (LOW): As a media processing tool, it ingests video data. While media decoders can have vulnerabilities, there is no mechanism here to interpolate video content back into the agent's control flow.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:00 PM