characteristic-voice
Warn
Audited by Snyk on Mar 19, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.80). SKILL.md's "Using a Character's Voice" workflow explicitly directs the agent to find a YouTube video and download subtitles/audio via yt-dlp and ffmpeg (public, user-generated content), read the subtitle to determine timestamps, and pass the downloaded audio as --ref-audio to the Noiz backend—thereby ingesting untrusted third‑party content that can influence tool behavior.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata