video-analysis

Pass

Audited by Gen Agent Trust Hub on Oct 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install official packages from the vendor's own repository and interact with their API.
  • Fetches resources and performs operations via the official vendor domain: api.sonilo.com.
  • Recommends installation of vendor-owned libraries: sonilo and sonilo-cli via pip and npm.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (videos) to generate text prompts for downstream tools, creating a potential injection surface.
  • Ingestion points: Reads external video files provided via video_path or video_url in the analyze_video tool (SKILL.md).
  • Boundary markers: None identified. The skill instructions suggest passing the output variations[i].prompt verbatim to subsequent tools.
  • Capability inventory: The skill output is intended to be used as input for subsequent generation tools like video_to_music or video_to_sfx (SKILL.md).
  • Sanitization: Absent. While there are mentions of copyright filters for output, there are no mechanisms described to sanitize the generated prompts against instructions embedded in the source video metadata or content.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 14, 2026, 08:39 PM