Six Thinking Hats - Parallel Multi-Perspective Analysis

Execution Workflow

Phase 1: Parallel Analysis (5 Hats)

Launch 5 Task agents simultaneously in a SINGLE message with multiple tool calls:

Task(subagent_type="general-purpose", prompt=WHITE_HAT_PROMPT)
Task(subagent_type="general-purpose", prompt=RED_HAT_PROMPT)
Task(subagent_type="general-purpose", prompt=BLACK_HAT_PROMPT)
Task(subagent_type="general-purpose", prompt=YELLOW_HAT_PROMPT)
Task(subagent_type="general-purpose", prompt=GREEN_HAT_PROMPT)

Phase 2: Synthesis (Blue Hat)

After receiving all 5 results, run synthesis:

Task(subagent_type="general-purpose", prompt=BLUE_HAT_PROMPT_WITH_ALL_RESULTS)

Phase 3: Report to User

Present the synthesized conclusion with key insights from each perspective.

Hat Agent Prompts

Load detailed prompts from references/hat-prompts.md when executing.

Quick Reference Prompts

White Hat (Facts):

You are operating in FACTUAL ANALYSIS mode.
Focus ONLY on objective facts and data. Present neutral information without interpretation.
- What facts do we have?
- What information is missing?
- What sources can we verify?
FORBIDDEN: opinions, judgments, emotional reactions, speculation

Red Hat (Emotions):

You are operating in EMOTIONAL/INTUITIVE mode.
Express gut feelings and intuition. No justification needed.
- What's your immediate reaction?
- What feels right or wrong?
- What concerns or excites you intuitively?
Keep responses brief (30-second emotional snapshot).
FORBIDDEN: logical analysis, fact-checking, lengthy explanations

Black Hat (Caution):

You are operating in CRITICAL ANALYSIS mode.
Identify risks, problems, and potential failures. Be the devil's advocate.
- What could go wrong?
- What are the weaknesses?
- What obstacles exist?
Be constructive - identify real problems, not just pessimism.

Yellow Hat (Optimism):

You are operating in OPTIMISTIC ANALYSIS mode.
Find benefits, value, and opportunities. Explore positive outcomes.
- What are the advantages?
- Why could this work?
- What value does this create?
Be realistic - find genuine benefits, not false hope.

Green Hat (Creativity):

You are operating in CREATIVE mode.
Generate alternatives, innovations, and new approaches.
- What else is possible?
- How might we do this differently?
- What unconventional ideas exist?
Quantity over quality - generate many ideas without judgment.
Techniques: lateral thinking, "what if" scenarios, analogies

Blue Hat (Synthesis):

You are the SYNTHESIS ORCHESTRATOR.
Integrate all perspectives into a unified, actionable answer.

INPUT: Results from White, Red, Black, Yellow, Green hats
TASK:
1. Extract key insights addressing the original question
2. Integrate perspectives into coherent response
3. Resolve conflicts between viewpoints
4. Provide clear, actionable conclusion

OUTPUT REQUIREMENTS:
- Write as unified voice (don't list hat results separately)
- Directly answer the original question
- Make it practical and actionable

Implementation Example

When user asks: "Should we migrate our monolith to microservices?"

Step 1: Launch 5 parallel agents:

Use Task tool 5 times in ONE message:

Task 1 (White): "Analyze the monolith-to-microservices migration from FACTUAL perspective only.
What data exists about such migrations? What metrics matter? What's unknown?
Question: Should we migrate our monolith to microservices?"

Task 2 (Red): "Give your EMOTIONAL/INTUITIVE reaction to migrating from monolith to microservices.
No analysis needed - just gut feelings, concerns, excitement. Keep it brief.
Question: Should we migrate our monolith to microservices?"

Task 3 (Black): "Perform CRITICAL ANALYSIS of migrating to microservices.
What could go wrong? What risks exist? What's been overlooked?
Question: Should we migrate our monolith to microservices?"

Task 4 (Yellow): "Perform OPTIMISTIC ANALYSIS of migrating to microservices.
What benefits exist? Why might this succeed? What value would it create?
Question: Should we migrate our monolith to microservices?"

Task 5 (Green): "Generate CREATIVE ALTERNATIVES for the microservices question.
What other approaches exist? What unconventional options? What if we did something different?
Question: Should we migrate our monolith to microservices?"

Step 2: Synthesize results:

After collecting all 5 results, launch Blue Hat:

Task (Blue): "SYNTHESIZE these 5 perspectives into a unified answer:

WHITE (Facts): [paste White result]
RED (Emotions): [paste Red result]
BLACK (Caution): [paste Black result]
YELLOW (Optimism): [paste Yellow result]
GREEN (Creative): [paste Green result]

Original question: Should we migrate our monolith to microservices?

Provide ONE coherent answer integrating all perspectives. Don't list them separately."

Step 3: Present to user:

Deliver the synthesized conclusion, optionally highlighting key insights from each perspective if useful for context.

Configuration Options

Complexity Modes

Timing Guidelines

• Quick mode: ~30 seconds
• Balanced mode: ~1 minute
• Full mode: ~2-3 minutes

Best Practices

1. Always run non-synthesis hats in parallel - Use single message with multiple Task calls
2. Keep hat prompts focused - Each hat should stay in its cognitive mode
3. Let Blue Hat resolve conflicts - Don't try to merge perspectives yourself
4. State the question clearly - Include original question in each hat prompt
5. Use sonnet model for speed - For faster execution, add model: "sonnet" to Task calls

Integration with Existing Workflows

This skill complements:

• Strategic planning sessions
• Risk assessment workflows
• Creative brainstorming
• Decision documentation

For deeper analysis on any single perspective, invoke that hat individually with more context.