What problem does it solve? Long Seedance 2.0 prompts often contain filler language that dilutes the production signal, but naive trimming can delete continuity locks, reference tags, or safety constraints that keep multi-clip sequences coherent. This Skill compresses prompts to 30-100 words (or compact Chinese) without losing the details that matter. ## Core Features & Use Cases - Priority-Based Compression: Preserves reference tags, subject identity, action endpoints, camera moves, light sources, audio cues, and safety constraints in a defined order while deleting generic adjectives and duplicate style labels. - Sequence-State Awareness: Retains continuity locks, actual opening state, completed beat exclusions, and reserved future beats so compressed prompts never replay finished action or leak upcoming beats. - Compact Templates: Provides ready-made short templates for T2V, I2V, V2V, and Chinese-language prompts. - Use Case: A user has a 300-word multi-clip Seedance prompt and needs a tight Chinese version for generation; the Skill returns a compact prompt that keeps @Image1 reference roles and the continuity constraint intact. ## Quick Start Compress my Seedance prompt into a short Chinese version under 100 words while keeping the reference tags and continuity constraints.
