---

name: symbolic-generation-structure description: "Route GETMusic, MuseCoco, Museformer, MeloForm, and EmoGen symbolic music generation workflows for any-track generation, text control, long-structure modeling, form refinement, and emotion control." disable-model-invocation: true metadata: disco-role: operating license: MIT

Symbolic Generation Structure

Use this sub-skill for Muzic projects that generate or refine symbolic music without lyric generation or retrieval: GETMusic, MuseCoco, Museformer, MeloForm, and EmoGen.

Start here

• Read references/getmusic.md for GETMusic track generation, position infilling, preprocessing, and training layout.
• Read references/musecoco.md for the two-stage text-to-attribute and attribute-to-music pipeline.
• Read references/structure-emotion-generation.md for Museformer, MeloForm, and EmoGen.
• Read references/troubleshooting.md for checkpoint, Fairseq, Triton, Java, jSymbolic, and prompt-format failures.
• Use scripts/validate_getmusic_request.py to check GETMusic track letters and position grammar before a generation run.
• Use scripts/plan_musecoco_pipeline.py to plan the MuseCoco stage-1 / stage-2 artifact handoff.

Route boundaries

• Understanding / retrieval workflows belong in music-understanding-retrieval.
• Lyric and songwriting workflows belong in lyric-melody-songwriting.
• MusicAgent orchestration belongs in music-agent-workflows.

Operating notes

• Keep this skill focused on artifact order, prompt grammar, and workflow boundaries.
• Treat training and inference recipes as checkpoint- and dataset-dependent unless the bundled references say otherwise.
• Validate input structure first, then pick the smallest matching workflow.
• When a request crosses into another Muzic family, route there instead of stretching the current recipe.