The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the subprocess module to orchestrate a wide range of external tools for video and audio processing, including ffmpeg, ffprobe, git, and the GitHub CLI (gh).
These tools are used for tasks like probing media files, joining video chunks, managing git repositories, and triggering cloud workflows.
[DYNAMIC_EXECUTION]: Dynamic Python code is generated and executed as part of the video processing pipeline.
In scripts/edit.py and scripts/modal_render.py, a multiline Python script defined in a JOIN constant is passed to python -c for final video assembly.
scripts/modal_render.py utilizes the Modal framework to serialize and execute local logic on remote cloud containers.
[REMOTE_CODE_EXECUTION]: The skill offloads computationally intensive rendering and processing tasks to cloud environments.
It triggers GitHub Actions workflows and Modal function calls to perform parallel video rendering.
Machine learning models for face detection and background matting are downloaded from external sources at runtime.
[EXTERNAL_DOWNLOADS]: The skill fetches assets, models, and configuration from various external services.
scripts/capture.mjs fetches icons and site logos from Simple Icons, jsDelivr, and Google APIs.
scripts/source.py uses yt-dlp to retrieve video sections and metadata from YouTube.
The skill automatically manages the installation of the Remotion renderer and Node.js dependencies.
[DATA_EXFILTRATION]: Local media files and project metadata are transmitted to external services for cloud processing.
Video footage is uploaded to user-controlled GitHub repositories and Modal Volumes for rendering.
[CREDENTIALS_UNSAFE]: The skill manages and utilizes third-party API keys and cloud credentials.
It retrieves secrets from ~/.aws/credentials, ~/.modal.toml, and .env files to authenticate sessions with ElevenLabs, AWS, and Modal.

style-edit