wtf is video-skills-toolkit?

Video Skills Toolkit packages three common short video production workflows into reusable templates and scripts, meant to be used as skills by an AI coding agent rather than as a single finished video project. It is not a video itself, but a set of starting points for building videos. The first piece, talking-head-remotion, is a Remotion project template for talking head style videos, the kind where a narrator speaks over a screen recording or graphics. It sets up a 16:9 studio look with a chapter progress bar at the top, a circular picture-in-picture frame in the bottom right corner, a safe area reserved for subtitles, and planning notes for sound effects and background music. The template deliberately starts with an empty middle section so you fill in the actual visuals scene by scene once you have your script and narration ready. The second piece, sketch-story-remotion, is aimed at simple hand drawn explainer style videos, using a cream colored paper background, stick figures, a small cartoon AI mascot, hand drawn outlines, arrows, and circle highlight animations. Every visual element is drawn through code rather than relying on image assets, so scenes can be generated and adjusted programmatically. The third piece, audio-to-subtitles, converts an audio or video file into subtitles and a full transcript, outputting SRT, VTT, or JSON formats, and it can also work from text-to-speech audio before converting that into subtitles. This part calls external services to do the actual transcription, so you need to supply your own API credentials for services such as MediaKit, R2, or MiniMax. No real API keys are included in the repository. Getting started with each template involves either running a Python scaffolding script or copying a template folder, then installing dependencies with npm and rendering a still frame or preview through Remotion's command line tools to confirm the setup works before building out a full video.