The Agent Skills Directory

Safe Command Execution: The skill utilizes ffmpeg and ffprobe for essential video processing tasks like trimming, scaling, and metadata inspection. These tools are invoked using subprocess.run with argument lists rather than shell strings, which is a recommended practice to prevent command injection.
Robust Credential Protection: The scripts demonstrate a high level of security awareness regarding API keys. The sanitize_error functions in generate_video.py and upload_file.py include extensive regex-based redaction to ensure that GEMINI_API_KEY values, OAuth tokens, and sensitive query parameters are removed from any error messages or logs before being displayed.
Trusted Resource Usage: All network communications and SDK dependencies are associated with official Google Gemini services. The skill uses the official google-genai Python SDK for model interactions and file management.
Data Handling Considerations: As an AI-powered media processing tool, the skill ingests user prompts and local media files. While this represents a potential surface for indirect prompt injection, the risk is mitigated by the underlying model's safety guardrails and the skill's specific focus on media generation rather than general-purpose system administration.

gemini-omni-flash-api