What problem does it solve? Rough ideas, abstract feelings, and over-specified shot scripts often produce weak results from text-to-video models. This Skill judges whether an input suits the Vibe Creating style and, when it does, distills it into a generation-ready video prompt while preserving user-specified dialogue, voiceover, music, and sound effects. ## Core Features & Use Cases - Judgment-first routing: Scores each input on scenario fit, expression form, and information density, then picks one of six actions (pass-through, light cleanup, direct rewrite, ask first, keep as-is, optional VC version) instead of blindly rewriting. - Constraint preservation: Keeps user-written dialogue, VO, music, SFX, shot numbers, and structure verbatim; technical camera parameters are translated into natural visual intent only when not explicitly protected. - Six rewrite modes: Narrative, emotional, memory, stream-of-consciousness, multi-shot experience, and mixed purification, selected by the input's dominant factor. - Use Case: A user pastes a cluttered prompt full of focal lengths and dolly distances for a nostalgic memory clip. The Skill strips the low-value technical control, keeps the emotional beats, and returns one clean prompt ready for Sora or Kling. ## Quick Start Ask the assistant to rewrite or vibe-ify your rough video idea or over-detailed shot script into a clean text-to-video prompt for a model like Sora or Kling.
