video-analyzer

Warn

Audited by Gen Agent Trust Hub on Feb 24, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands using yt-dlp, ffmpeg, and curl. It directly interpolates user-controlled variables like the video URL and prompt text into these commands, creating a risk of command injection if the inputs contain shell metacharacters. The skill also includes a command that accesses local browser cookies (--cookies-from-browser chrome) for Bilibili authentication.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes yt-dlp to download media from external URLs. This involves fetching untrusted content from the internet and saving it to the local system.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection from the processed video content. Ingestion points: Video URL in SKILL.md. Boundary markers: Absent in shell commands and API prompts. Capability inventory: yt-dlp, ffmpeg, curl, rm. Sanitization: Absent.
  • [SAFE]: The skill transmits data to the Volcengine API (ark.cn-beijing.volces.com), which is a well-known service. This behavior is consistent with the skill's primary function.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 24, 2026, 02:42 PM