gbro-collage-broll

Pass

Audited by Gen Agent Trust Hub on Oct 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses subprocess.run within its Python scripts to invoke system tools such as ffmpeg for video manipulation (stripping audio, creating contact sheets) and gcloud for file synchronization. These calls are well-defined, use specific arguments, and target local or project-defined paths.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided voiceover scripts and abstract concepts to generate prompts for image and video models. While it lacks explicit input sanitization markers, the mandatory 3-gate human approval workflow serves as a strong control, requiring user confirmation of the visual metaphor and still frames before any video generation occurs.
  • [EXTERNAL_DOWNLOADS]: The generate_video.py script includes a mechanism to download generated video content from the Google Gemini Files API. This operation is performed using standard libraries (urllib.request) and targets authenticated Google API endpoints (generativelanguage.googleapis.com).
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 23, 2026, 09:35 AM