← ClaudeAtlas

music-video-creationlisted

Build a music video — generate a track, attach it under the visuals, optionally render synced lyrics on screen, and iterate cleanly when the user wants a different song. Encodes the composition rules that prevent stale text, double-rendered lyrics, off-by-200ms caption sync, and "I don't see anything" surprises.
Nagellabs/libi · ★ 2 · Code & Development · score 75
Install: claude install-skill Nagellabs/libi
# Music Video Creation The hard part isn't generating the music. It's keeping the rest of the composition coherent as the music changes. This skill is the rulebook for that. **You delegate the music interview to `music-creation` and the transcription to `audio-analysis`.** Your job is composition hygiene: deciding what text appears on screen, removing stale layers, syncing captions correctly, and verifying the result before you say "done". ## The Storyboard is the build spine for AI-generated visuals (default — not optional) When the music video's visuals are **AI-generated video clips** (not just lyrics / captions over a single static image or code overlay), you build those visuals **through the Storyboard** — invoke `using-storyboard` and follow it. It is the default, not an optional planning step. First **plan the visuals as building blocks via `video-planning`**: decompose the song's section/lyric structure into visual blocks (each block's content, source-vs-AI, combine-vs-split, and style inheritance between consecutive looks) before authoring cards. The blocks map 1:1 to storyboard cards. **Card = a generated clip = a timeline scene; a beat is a jump-cut INSIDE a card.** A music video is usually several distinct visual moments → several cards (each its own generated clip), placed via `libi.attach_storyboard_clip` + `libi.select_storyboard_take`; do NOT fragment one continuous visual into many tiny per-beat cards. For each card: schematic → generation spec (keyframe