What problem does it solve? Writing effective prompts for AI music generation is difficult: brief descriptions lack the detail models need, and users struggle to translate musical intent into structured, section-aware instructions. This Skill turns a short music description and optional tagged lyrics into a professional MiniMax Music 3 structured caption ready for the generation API. ## Core Features & Use Cases - Structured Caption Generation: Produces captions with exactly three sections — Global Metadata (genre, tempo, emotional progression, production profile), Vocal Details (timbre, register, delivery, harmonies), and a section-by-section Arrangement timeline. - Progressive Template Retrieval: Routes genre and mood cues through a genre router to one of 18 style families, selects up to three reference templates (Foundation, Modifier, Arrangement) from a library of 1,000 text templates, and synthesizes an original caption without copying template content. - Lyric Tag Preservation: Treats bracketed lyric tags like [Verse] and [Chorus] as executable structural directives while never quoting or reproducing lyric content. - Use Case: A user provides "a warm acoustic pop song with intimate female vocals and a gradual emotional build" plus tagged lyrics, and receives a complete structured caption to pass directly as the instructions field of the MiniMax Music 3 generation API. ## Quick Start Use $music-caption-rewriter to turn my music description and tagged lyrics into a structured Music 3.0 caption with global metadata, vocal details, and arrangement.
