The Agent Skills Directory

[COMMAND_EXECUTION]: The skill frequently uses the subprocess.run module in multiple scripts (including extract_shorts.py, render_vertical.py, and prepare_transcript.py) to execute standard video processing utilities like ffmpeg and ffprobe. These calls are implemented using list-based arguments rather than shell strings, which effectively mitigates the risk of shell injection.
[SAFE]: The skill incorporates advanced integrity checks by calculating and verifying SHA-256 fingerprints for all critical inputs and intermediate artifacts (video files, transcripts, and timeline metadata). This prevents unauthorized data tampering or 'file-swapping' attacks between the planning and rendering phases.
[SAFE]: A sophisticated 'Review Gate' system is implemented in review_gate.py, requiring explicit human or delegated agent approval through immutable, hash-bound HTML pages before final video extraction or rendering can occur.
[SAFE]: The skill ensures that all user-controlled or agent-authored metadata displayed in the review pages is properly sanitized using HTML escaping (found in preview.py and the HTML templates), preventing potential Cross-Site Scripting (XSS) vulnerabilities.
[SAFE]: All external tools (ffmpeg, ffprobe) and internal script dependencies are validated for existence and integrity before execution.

video-to-shorts