veo-use

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • SAFE (SAFE): No malicious behavior, obfuscation, or unauthorized data access patterns were found. The skill follows best practices by using environment variables for authentication and leveraging official libraries.\n- Indirect Prompt Injection (LOW): The scripts ingest user-provided prompts and media files which are passed to an external AI model. While this represents a standard attack surface for LLM-based tools, there is no evidence of malicious exploitation within the skill itself.\n
  • Ingestion points: Command-line arguments such as --prompt, --image, --video, and --mask in edit_video.py, extend_video.py, image_to_video.py, reference_to_video.py, and text_to_video.py.\n
  • Boundary markers (absent): No delimiters or specific 'ignore embedded instructions' warnings are applied to the user-provided prompts before being sent to the API.\n
  • Capability inventory: The skill possesses the capability to make network requests to Google GenAI APIs (client.models.generate_videos) and write files to the local filesystem (vid.video.save).\n
  • Sanitization (absent): No explicit sanitization, escaping, or validation of the input prompt is performed by the scripts prior to API submission.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:43 PM