waker240/fullvideoproductionskill

This is a packaged skill for the Cursor AI coding tool, built to help turn a written video script into an actual animated explainer video. The README describes it as a translation layer that takes a script you already wrote, including your argument, characters, and pacing, and turns it into concrete visual decisions for a code based animated video built with Remotion, a tool for making videos out of code. The package bundles the skill's own documentation, a working reference implementation of the video building blocks it describes, and a small server that gives the AI agent tools called generate_image and generate_video. It is built to work with plain Remotion rather than any specific proprietary video framework, and everything needed to actually use the skill, not just read about it, lives inside this one folder. The package deliberately leaves out the earlier step of writing the script itself, since it assumes you already have one, and it does not include a full ready to run Remotion project, only the reusable pieces you drop into your own project. Setting it up involves several steps. You copy the skill folders into Cursor's skill directory so Cursor can find and call them. You install and start a small local server that gives the agent image and video generation tools, configured with your own API key for routing to an image or video model. You copy two folders of reusable code, called the canvas layer and the cinema layer, directly into your own Remotion project's source folder. You also copy two rule files into your project that Cursor applies automatically: one describing a full pipeline for turning raw scene audio into a finished, timed video with subtitles, and another describing a subtitle system built to highlight key phrases in sync with the argument being made. The package also includes a folder of reference notes distilled from twelve books about visual storytelling, covering things like pacing, shot composition, and film editing, meant to explain the reasoning behind the skill's visual choices rather than to be read once and set aside. A separate guide walks through generating video thumbnail images using an image generation model, including sizing rules and several named visual styles to choose from.