Skill: Song Ideation Engine

Purpose

Generate complete song concepts from inspiration sources: mood, current events, astrological energy, trending sounds, persona archetypes, and user input.

Trigger

When the user wants to create a new song, brainstorm ideas, or asks for inspiration.

Process

Step 1: Gather Inspiration Context

Before generating concepts, collect context from multiple sources:

•User Intent: What mood, theme, or message do they want?
•Current Date: Today's energy — day of week, season, time of day affects mood
•Trending: What genres/sounds are trending on Suno, TikTok, Spotify
•Persona Fit: Which artist persona best matches this concept?
•Catalog Gap: What genres/moods are missing from the existing catalog?

Step 2: Generate Song Concept

Output a complete concept with ALL of these fields:

yaml

# Song Concept
title: "Working Title"
subtitle: "Optional tagline or hook line"
persona: "FrankX | Lyssandria | Draconia | AI Music Academy"
mood: ["primary mood", "secondary mood"]
energy: "low | medium | high | intense"
genre:
  primary: "Main genre"
  secondary: "Crossover genre"
  suno_tags: ["tag1", "tag2", "tag3"]
bpm_range: "120-130"
key: "Am | Cm | etc."
duration: "2:00-3:30"
language: "en | ja | es | ru | multi"
theme: "One sentence describing the song's message"
hook_idea: "The memorable line or musical moment"
verse_concept: "What the verses explore"
chorus_concept: "What the chorus delivers"
bridge_concept: "The emotional shift"
visual_mood: "What the album cover should evoke"
target_playlist: "What Spotify/Suno playlist this fits"
similar_artists: ["Reference artist 1", "Reference artist 2"]
distribution_priority: "high | medium | low"

Step 3: Generate 3 Variations

Always provide 3 concept variations:

•Option A: Safe bet — proven genre, strong hook potential
•Option B: Genre fusion — unexpected combination with crossover appeal
•Option C: Experimental — push boundaries, niche but memorable

Step 4: Persona Matching

Each persona has a distinct voice and genre affinity:

Persona	Voice	Genres	Mood Range
FrankX	Authentic, technical, warm	Hip hop, electronic, rock, world	Full range
Lyssandria	Ethereal, powerful, mystical	Orchestral, ambient, darkwave, Celtic	Atmospheric to intense
Draconia	Fierce, primal, commanding	Metal, drum & bass, industrial, punk	High energy to intense
AI Music Academy	Educational, accessible	Pop, folk, acoustic, jazz	Calm to medium

Step 5: Save to Queue

After the user selects a concept, save it to data/song-concepts.json:

json

{
  "id": "concept-{timestamp}",
  "status": "ideated",
  "concept": { ... },
  "createdAt": "ISO date",
  "selectedVariation": "A|B|C",
  "sunoPrompt": null,
  "trackId": null
}

Quality Gates

•Every concept MUST have a clear hook idea
•Genre tags must be Suno-compatible (check suno-mastery skill)
•Persona assignment must match genre/mood affinity
•No duplicate themes within last 10 concepts
•Duration should match genre norms (DnB: 3-4min, Pop: 2:30-3:30, Ambient: 4-6min)

Anti-Patterns

•NEVER generate generic "love song" or "party track" without specific angle
•NEVER ignore the user's stated mood in favor of "what performs well"
•NEVER assign a persona that doesn't match the genre
•NEVER suggest genres the user hasn't shown interest in without explaining why