What problem does it solve? Image generators draw static poses but cannot reliably imagine mid-motion frames, so hand-requested animation keyframes often break continuity and fail to loop. This Skill solves that by animating a base frame with a video model and harvesting real motion frames that cycle correctly. ## Core Features & Use Cases - Video-first frame harvest: Animate a neutral base pose with image-to-video, then extract frames that capture distinct motion phases and close the loop. - Pipeline integration: Delegates generation and chroma-key postprocessing to the video2dsprite and generate2dsprite skills using a #FF00FF magenta background. - Motion laws and flip test: Verifies loop closure, limb continuity, physics readability, and energy level before frames ship into a game. - Use Case: You need a side-view walk cycle for a knight character. Generate a base still on magenta, animate it walking in place, harvest 8-12 frames, run the flip test, and package a zero-padded sequence or sprite sheet with a stated fps. ## Quick Start Ask the AI to create a looping walk cycle sprite sheet for your character using the video2dsprite pipeline and verify it with the flip test.
