video-downloader

Pass

Audited by Gen Agent Trust Hub on Jul 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The scripts/download_video.py script executes the yt-dlp tool using subprocess.run. The implementation uses argument lists which prevents shell injection from malicious URLs.
  • [EXTERNAL_DOWNLOADS]: The skill downloads video and audio content from external hosting platforms (YouTube, Instagram, etc.) as its primary function.
  • [PROMPT_INJECTION]: Analysis of the indirect prompt injection surface identifies that video metadata is fetched from external URLs. \n
  • Ingestion points: Video metadata (title, uploader) is retrieved from external sources and displayed to the agent in scripts/download_video.py.\n
  • Boundary markers: Present in SKILL.md, instructing the agent to treat scraped content as untrusted data and ignore embedded instructions.\n
  • Capability inventory: Includes subprocess execution and filesystem writes in scripts/download_video.py.\n
  • Sanitization: Metadata is printed directly to stdout, but SKILL.md provides sufficient context to prevent obedience to embedded instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 30, 2026, 03:50 PM