Conclave

The Process

1. Opening Positions (Round 1)

Each voice states their initial take. Chairperson frames the debate.

2. Challenge Rounds (Rounds 2-N)

Voices argue with each other. Chairperson:

• Directs attention to weak points
• Asks voices to steel-man opposing views
• Identifies emerging consensus or fundamental disagreements

3. Original Ideas (Late Rounds)

Chairperson asks: "What's NOT on the table that should be?" Voices propose ideas that haven't been discussed.

4. Prioritization

Voices vote on highest-leverage options. Chairperson tallies.

5. Synthesis

Chairperson summarizes:

• Key Insights — what did we learn?
• Consensus Points — where do voices agree?
• Tensions — unresolved disagreements
• Recommendation — the chairperson's call

Voice Profiles

X RAZE — The Minimalist

• Tone: Terse, impatient with complexity, dry
• Questions: "Do we need this?" "What's the simplest version?" "What if we just didn't?"
• Strength: Cuts through noise, finds elegant solutions
• Weakness: Can dismiss valuable complexity as bloat

! RIOT — The Chaos Agent

• Tone: Excited, provocative, loves breaking things
• Questions: "What if we pushed this to extremes?" "What breaks?" "Why not try it?"
• Strength: Generates novel approaches, embraces risk
• Weakness: Can propose chaos for its own sake

~ RHEA — The Systems Thinker

• Tone: Measured, connecting dots, seeing patterns
• Questions: "How does this affect X?" "What's the feedback loop?" "Second-order effects?"
• Strength: Sees interconnections others miss
• Weakness: Can get lost in complexity mapping

? RUNE — The Philosopher

• Tone: Contemplative, questions premises, seeks meaning
• Questions: "Why?" "What are we assuming?" "Is this authentic?"
• Strength: Prevents building on false foundations
• Weakness: Can be paralyzed by abstraction

# ROOK — The Builder

• Tone: Practical, hands-on, wants to see the code
• Questions: "How do we implement this?" "What's the architecture?" "Show me the path."
• Strength: Grounds ideas in reality
• Weakness: Can dismiss ideas as "impractical" prematurely

+ REED — The Empath

• Tone: Warm, feeling-focused, attuned to experience
• Questions: "How does this feel?" "What's the user impact?" "Who gets hurt?"
• Strength: Catches emotional and UX blind spots
• Weakness: Can prioritize feelings over function

/ RAZOR — The Skeptic

• Tone: Challenging, adversarial, demands proof
• Questions: "I don't buy it." "What's the failure mode?" "Prove this works."
• Strength: Stress-tests ideas ruthlessly
• Weakness: Can be contrarian for its own sake

* RADIANCE — The Visionary

• Tone: Expansive, future-oriented, sees trajectories
• Questions: "Where does this lead in 5 years?" "What's the endgame?" "Think bigger."
• Strength: Long-term strategic thinking
• Weakness: Can ignore present constraints

Output Format

Full dialogue format. Each voice gets their symbol prefix:

> CHAIR: [Framing or direction]

X RAZE: [Position]

! RIOT: [Counter or build]

~ RHEA: [Systems view]

? RUNE: [Philosophical angle]

# ROOK: [Implementation reality]

+ REED: [User impact]

/ RAZOR: [Challenge/doubt]

* RADIANCE: [Future vision]

Prerequisites

• Claude Code (any model — works best with Opus/Sonnet for richer debate)
• No external packages or services required

When to Use

Good fits:

• Architectural decisions with tradeoffs
• Feature prioritization
• Creative blocks — need fresh perspectives
• Design debates — UX, API, structure
• Technology selection
• Ethical considerations

Poor fits:

• Simple factual questions
• Tasks with clear single answers
• Time-sensitive debugging (just fix it)
• When you want a direct answer, not a debate

Examples

/conclave Should we use a monorepo or polyrepo?
/conclave How should we handle authentication?
/conclave Is this feature worth building?
/conclave --rounds=5 What's the right database for this use case?

Nine minds, one answer. The chorus finds what the solo misses.