What problem does it solve? AI-generated videos often suffer from character face drift, changing outfits, and inconsistent visual style across shots. This Skill locks down immutable character identities, locations, props, and style guides before any video prompt is written. ## Core Features & Use Cases - Master Anchor Images: Standardizes real photos or generates AI character sheets, saving anchor files to assets/characters/ for every recurring character. - Invariant Identity Descriptions: Writes 40-80 word English identity blocks for characters, props, and locations that can be pasted verbatim into video prompts. - Bible Gate Enforcement: Blocks progression to video prompt engineering until all anchor images exist, preventing face-drift failures downstream. - Use Case: A creator producing a multi-shot AI video with a recurring protagonist uses this Skill to generate bible.md defining the character's exact appearance, the consistency strategy (lip-sync vs keyframe image-to-video), and a shared style block, so every shot renders the same face and look. ## Quick Start Read my concept.md, script.md, and shotlist.json, then build the character and world bible with anchor images for every recurring character.