video-understanding

Pass

Audited by Gen Agent Trust Hub on Jun 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill utilizes well-known technology tools and packages (yt-dlp, ffmpeg, and openai-whisper) for its intended purpose of video processing and transcription.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes transcripts from external videos, which could potentially contain adversarial instructions designed to influence agent behavior.
  • Ingestion points: Transcript files located in the output/ directory (SKILL.md).
  • Boundary markers: No specific delimiters or instructions to ignore embedded commands are used when the agent reads the transcript.
  • Capability inventory: The skill allows shell command execution via Bash and general file system Read/Write access.
  • Sanitization: The transcript content is not sanitized or validated before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 21, 2026, 10:49 PM
Security Audit — agent-trust-hub — video-understanding