The Agent Skills Directory

[COMMAND_EXECUTION]: The Python scripts in the edit/ directory (01_transcribe.py, 03_apply_cuts.py, 04_speedup.py) execute external system commands to process video and audio.
Evidence: Use of subprocess.run() with argument lists to invoke ffmpeg, ffprobe, and whisper-cli safely.
Context: These operations are essential for the skill's purpose of transcribing and cutting video footage locally.
[EXTERNAL_DOWNLOADS]: The skill's documentation directs users to install specific external tools and skills.
Evidence: Recommendations to install heygen-com/hyperframes and whisper-cpp via package managers.
Context: These are well-known dependencies for the video rendering engine and transcription components.
[SAFE]: The production workflow is explicitly designed with human oversight.
Evidence: SKILL.md enforces two mandatory approval gates where the agent must pause and wait for user verification before proceeding to storyboard assets or final rendering.
Context: This design pattern prevents the agent from performing major actions or consuming resources without explicit user consent.

short-form-studio