youtube-downloader

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • EXTERNAL_DOWNLOADS (MEDIUM): The skill performs automatic runtime installation of the 'yt-dlp' package from PyPI if the binary is not found on the host system.\n
  • Evidence: The check_yt_dlp function in scripts/download_video.py executes pip install --break-system-packages yt-dlp. While PyPI is a standard registry, runtime installation of external code without version pinning or integrity checks is a security risk.\n- COMMAND_EXECUTION (SAFE): The script uses subprocess.run with a list of arguments for all external calls, which follows security best practices to prevent shell injection.\n- PROMPT_INJECTION (LOW): The skill is susceptible to indirect prompt injection because it ingests untrusted metadata (titles, uploader names) from YouTube and reflects it into the agent's output.\n
  • Ingestion points: Video metadata fetched via yt-dlp --dump-json in scripts/download_video.py.\n
  • Boundary markers: Absent; the script prints metadata directly to the console without delimiters.\n
  • Capability inventory: The skill can execute subprocesses (yt-dlp) and write files to /mnt/user-data/outputs/.\n
  • Sanitization: Absent; the script relies on the external tool's default behavior for filename safety.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:57 PM