Stats
Actions
Tags
From suno-composer
Compose Suno AI songs with creative or guided workflow, adaptive preferences, and session reflection
How this command is triggered — by the user, by Claude, or both
Slash command
/suno-composer:suno [theme, artist, mood, or album/variation/extend intent]This command is limited to the following tools:
The summary Claude sees in its command listing — used to decide when to auto-load this command
# Suno Song Composition Compose songs optimized for Suno AI v5. ## Step 0: Load Context ### Preferences 1. Check for preference files: `~/.claude/suno-composer.local.md` (global), `.claude/suno-composer.local.md` (project) 2. If both exist: merge (project sections override matching global sections, global fills gaps) 3. If neither exists: trigger First-Run Wizard ### First-Run Wizard **Trigger:** No preference file at either location. 1. Offer setup: "No preferences found. Quick setup? (~20 seconds)" - Options: "Yes" / "Skip" 2. If yes, ask 2-3 questions via AskUserQuestion: - ...
Compose songs optimized for Suno AI v5.
~/.claude/suno-composer.local.md (global), .claude/suno-composer.local.md (project)Trigger: No preference file at either location.
# Suno Composer Preferences
## Favorite Genres
## Preferred Vocal Styles
## Default Languages
Invoke the suno skill via the Skill tool.
| Signal | Mode | Context Load |
|---|---|---|
| Album/EP/multi-track intent ("album about...", "5-track EP", "collection of...") | Album | references/album-composition.md |
| Variation intent ("acoustic version", "remix of", "stripped down") | Variation | references/variation-patterns.md |
| Continuation intent ("sequel to", "what happens next", "prequel", "response song") | Extend | references/continuation-patterns.md |
| Artist reference ("like X", "in the style of X") | Vision-First | references/artist-profiles.md |
| Tier keyword (anisong, surface, mainstream, doujin, legacy) | Vision-First | references/jpop-tiers.md |
| Rich creative direction (theme + genre/mood) | Vision-First | — |
| Sparse or empty input | Guided | — |
| Ambiguous | Ask via AskUserQuestion | — |
If both tier and artist detected: tier provides base sound, artist provides specific refinements. Artist takes precedence for conflicts.
All modes follow this core pattern after gathering parameters:
./songs/[timestamp]-[slug]/ directory
song-N-[title-slug].md with Style Prompt, Lyrics, Specifications sections_index.md with track listingreferences/output-formats.md for full templatesDefault for rich input.
Parse creative direction and immediately envision concept, format (single/EP/album), sound, language, and vocals. Infer from context; ask only when genuinely ambiguous.
Present a vivid creative vision — imagery, not metadata:
I'm imagining "[Title]" — a [N]-track [format] about [vivid concept]:
1. "[Track]" — [emotional/narrative hook]
2. "[Track]" — [story progression or mood shift]
...
Sound: [genre feel], [tempo], [production]. [Language] lyrics, [vocal character].
Shall I compose this? Or adjust the direction?
Make specific artistic choices. Don't hedge with "could be X or Y."
Handle natural reactions: "darker", "fewer tracks", "make it Korean", "not quite, I meant..." — adjust and re-present until confirmed. Then proceed to Shared Workflow.
Triggered by empty or sparse input.
Ask via AskUserQuestion, combining where possible:
Smart defaults: if user provided partial info (e.g., "japanese ballad"), pre-fill known values and skip those questions.
Generate metadata previews per song (title, direction, genre, tempo, vocal, arc, hook concept). Then proceed to Shared Workflow.
Activated when user describes an album, EP, or multi-track collection.
Define album framework: thematic anchor, sonic palette, tempo range. Plan track sequence with roles (opener, journey, peak, descent, closer) and energy levels.
Preview format:
## Album: [Title]
Concept: [description] | Sonic Palette: [instruments, production, tempo range]
1. [Title] (Opener) — [genre], ~[BPM] — [theme]
2. [Title] (Journey) — [genre], ~[BPM] — [theme]
...
Then proceed to Shared Workflow. Write to album structure: [timestamp]-[album-name]/ with _album.md overview + numbered track files.
Activated when user requests a transformed version of an existing song.
Analyze source DNA (hook, theme, sonic elements). Preview each variation with key transformations from original.
Then proceed to Shared Workflow. Write to: [timestamp]-[source]-variations/ with _source.md + variation files.
Activated when user wants a narrative continuation of an existing song.
Extract source DNA and plan continuation based on type. Include planned callbacks (2-3 lyrical callbacks per song). Preview with connection description and callback concepts.
Then proceed to Shared Workflow. Maintain sonic DNA (tempo +/-15 BPM, related key, shared instruments). Write to: [timestamp]-[title]/ with _connection.md + song file + listening-order.md.
After composition, naturally reflect on patterns observed during the session. Skip for single-song sessions or brief interactions.
Reflect when multiple songs revealed consistent creative patterns (genre, mood, artist, mode, language preferences).
Offer conversationally — one specific observation, not a checklist:
"You referenced Aimer three times. Should I default to her style?"
If accepted: Ask global vs project, then append to .local.md in natural language. Add to existing sections or create new ones. Keep observations specific and actionable.
If no patterns emerged: End session without reflection.
npx claudepluginhub linaqruf/kana-code-plugins --plugin suno-composer/gen-audioGenerates an AI audio brief for voiceover or music production, with tool selection, timing cues for video sync, and synthetic voice compliance notes.
/generateCreates a voice profile by analyzing your writing samples via the voiceprint skill workflow, including introduction, setup, and multi-phase processing (~15 min).
/repurpose-talkTransforms a recorded talk or presentation into 10-30 repurposed content pieces (social posts, blog drafts, clips, etc.).