video-provider-capabilities

Pass

Audited by Gen Agent Trust Hub on Oct 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes ffprobe to verify the technical properties (codec, duration, dimensions) of downloaded video files and show_video for presenting generated content to the user for review. These are standard diagnostic and preview tools in a video production workflow.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, creating an attack surface for indirect instructions. It reads official API documentation (SKILL.md) and processes generated video files (references/execution-review.md).
  • Ingestion points: Provider API documentation pages and generated video files.
  • Boundary markers: No explicit delimiters are specified for the documentation content.
  • Capability inventory: Includes ffprobe and show_video calls, and file system writes for downloads.
  • Sanitization: No sanitization is mentioned for data extracted from external documentation or video file metadata.
  • [EXTERNAL_DOWNLOADS]: The skill involves fetching live API schemas and downloading video outputs from external providers. The references target established organizations and well-known services (Google, fal.ai) in the AI ecosystem.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 7, 2026, 09:42 AM