The Agent Skills Directory

Command Execution for Media Processing: The skill relies on system-level binaries ffmpeg and ffprobe to perform tasks such as video trimming, resolution scaling, and audio stream analysis. These operations are conducted through Python's subprocess.run using argument lists, which is a standard method for integrating media tools into automated workflows.
Active Credential Sanitization: A notable security feature within scripts/video/generate_video.py and scripts/upload_file.py is the sanitize_error function. This utility is specifically designed to detect and redact sensitive patterns, including API keys, OAuth tokens, and internal infrastructure URLs, from error logs and console outputs to prevent accidental data exposure.
Controlled Filesystem Interaction: The skill manages media assets by reading from and writing to local paths (defaulting to a media/ directory). This level of filesystem access is necessary for the skill's primary function of preparing and generating video files.
Vendor-Aligned Dependencies: The skill utilizes the official google-genai SDK and references documentation from Google's AI developer platforms. These dependencies are appropriate for the skill's author and the intended integration with the Gemini Omni model.

gemini-omni-flash-api