suno

The music-generation tool skill — target the brief, rights first, arrange with structure, curate + humanize, keep disclosure clean. The agent briefs (and can drive the API where connected); the human's ears judge; WoopSocial publishes the content the music scores. (Ships with tools/integrations/suno.md.)

The POV: a full song from a sentence — but the rights are the craft

Suno v5.5 makes the most natural vocals in the category, splits its own generations into clean stems, and now offers a real consistency stack (Personas → your own verified Voice → a Custom Model trained on tracks you own). But 2026 Suno is defined by its legal moment, and the top-1% operator works rights-first. (1) Paid plan before the first note: free-tier songs are non-commercial forever — upgrading never retroactively unlocks them. (2) Commercial rights ≠ ownership: the post-Warner terms grant paid users commercial rights while Suno's own docs concede purely-AI music isn't copyrightable — you can monetize, but you can barely stop theft of a raw track; human elements (original lyrics, a re-recorded part over stems, real DAW arrangement) strengthen the claim and reclassify the work as AI-assisted. (3) No indemnification while the lawsuits run: Warner settled and partnered (Nov 2025), but UMG and Sony are still suing toward a July 2026 fair-use hearing — so favor generic style language over artist references, and never clone an artist's voice (the opt-in program is the consent path, not prompting). (4) Archive now: pre-deal models face deprecation — WAVs, stems, MIDI out. And the social-stack truth: trending audio is native-only — Suno is the owned brand sound, not the trend sound.