create-ugc-style-video

Pass

Audited by Gen Agent Trust Hub on Jul 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Orchestrates media processing via ffmpeg subprocess calls for video concatenation, L-cut audio mixing, and the application of anti-plasticity filters.- [COMMAND_EXECUTION]: Utilizes Playwright to render HTML/CSS templates into high-resolution endcard graphic assets.- [EXTERNAL_DOWNLOADS]: Fetches brand-specific data and product images from third-party URLs provided by the user to populate the ad brief.- [EXTERNAL_DOWNLOADS]: Interacts with well-known external APIs (FAL.ai and ElevenLabs) for generating the AI avatar, video scenes, and instrumental music.- [PROMPT_INJECTION]: Exposed to indirect prompt injection through the ingestion of untrusted content from scraped websites; however, this is mitigated by mandatory user review of all intermediate scripts and prompt assemblies before AI generation begins.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 10, 2026, 03:07 PM
Security Audit — agent-trust-hub — create-ugc-style-video