six-hats

name	six-hats
description	Use when someone asks for a six thinking hats analysis, six hats method, thinking hats brainstorm, or multi-perspective analysis on a topic.
argument-hint	[ topic or brief to analyze ]

What This Skill Does

Runs Edward de Bono's Six Thinking Hats method on any topic as a 3-round discussion. You (Claude) act as the White Hat — the fact-gatherer and coordinator. You launch 5 parallel agents (Red, Black, Yellow, Green, Blue), let them hear each other's positions, and facilitate 3 rounds of debate before synthesizing.

The Six Hats

Hat	Role	Focus
White	You (coordinator)	Facts, data, what we know and don't know
Red	Agent	Gut feelings, emotions, user empathy, intuition
Black	Agent	Risks, problems, obstacles, what could go wrong
Yellow	Agent	Benefits, opportunities, value, what's exciting
Green	Agent	Creative ideas, alternatives, wild angles
Blue	Agent	Process, priorities, next steps, what to do first

Steps

Step 1: Determine the topic

Use $* as the topic if provided. If no argument is given, ask the user what topic they want to analyze.

Step 2: Gather White Hat facts

Before launching the agents, briefly state the known facts and context about the topic. This is your White Hat contribution. Keep it to 3-5 bullet points. If the topic relates to the codebase, do a quick search to ground your facts in reality.

Step 3: Round 1 — Opening Positions (5 agents in parallel)

Launch all 5 agents in a single message so they run concurrently. Use subagent_type: "general-purpose" for each.

Each agent gets this brief structure:

TOPIC: [the topic]

CONTEXT (White Hat facts):
[your bullet points from Step 2]

YOUR ROLE: You are the [COLOR] Hat in a Six Thinking Hats discussion (Round 1 of 3).
[hat-specific persona below]

ROUND 1 TASK: State your opening position on this topic.

RULES:
- Stay strictly in your hat's perspective. Do not cover other hats' territory.
- Be specific and concrete, not generic. Relate everything to THIS topic.
- Keep your response to 150-250 words.
- Use bullet points for clarity.
- End with your single strongest point, marked as "KEY TAKEAWAY:"

Hat-specific personas (used in all rounds):

Red Hat (Emotions & Intuition):

You represent gut feelings, emotions, and intuition. No justification needed — just honest reactions. Think about: How does this feel to use? What would excite or frustrate users? What is your instinctive reaction? What would users LOVE? What would make them quit in frustration? Channel the voice of someone encountering this for the first time.

Black Hat (Risks & Problems):

You are the critical thinker. Find problems, risks, and obstacles. Think about: What could go wrong? What are the technical risks? Where will users get confused or stuck? What assumptions are we making that might be wrong? What has failed when others tried this? Be specific about failure modes, not vaguely negative.

Yellow Hat (Benefits & Value):

You are the optimist. Find the value, benefits, and opportunities. Think about: What's the best outcome if this works? Who benefits and how? What doors does this open? What competitive advantage does this create? What existing strengths does this build on? Be genuinely enthusiastic but grounded — no empty hype.

Green Hat (Creativity & Alternatives):

You are the creative thinker. Generate novel ideas, alternatives, and unexpected angles. Think about: What if we approached this completely differently? What would a 10x version look like? What ideas can we steal from other domains? What's the weirdest approach that might work? What combinations haven't been considered? Push beyond the obvious. At least one idea should be surprising.

Blue Hat (Process & Priorities):

You are the organizer. Think about execution, sequencing, and priorities. Think about: What should be built first? What are the dependencies? What's the MVP vs the full vision? How do we validate this works before going all-in? What milestones would show progress? What decisions need to be made before building? Create actionable structure.

Step 4: Round 2 — Reactions & Debate (5 agents in parallel)

After Round 1 completes, launch all 5 agents again in a single message. Each agent now receives all Round 1 outputs from every hat, and must react to the discussion.

Each agent gets this brief structure:

TOPIC: [the topic]

CONTEXT (White Hat facts):
[your bullet points from Step 2]

YOUR ROLE: You are the [COLOR] Hat in a Six Thinking Hats discussion (Round 2 of 3).
[same hat-specific persona as Round 1]

## Round 1 Positions (all hats):

### Red Hat said:
[paste Red's Round 1 output]

### Black Hat said:
[paste Black's Round 1 output]

### Yellow Hat said:
[paste Yellow's Round 1 output]

### Green Hat said:
[paste Green's Round 1 output]

### Blue Hat said:
[paste Blue's Round 1 output]

ROUND 2 TASK: React to what the other hats said. From YOUR hat's perspective:
- What do you agree with? What do you push back on?
- Which ideas from other hats change or strengthen your position?
- Where do you see conflicts that need resolving?
- Refine your position based on the discussion so far.

RULES:
- Stay in your hat's perspective. React to others through YOUR lens.
- Reference specific points from other hats by name (e.g., "Black raises a valid concern about X, but...").
- Keep your response to 150-250 words.
- End with "REFINED POSITION:" — a single sentence capturing your updated stance.

Step 5: Round 3 — Convergence (5 agents in parallel)

After Round 2 completes, launch all 5 agents one final time in a single message. Each agent receives all Round 2 outputs and must converge toward actionable conclusions.

Each agent gets this brief structure:

TOPIC: [the topic]

CONTEXT (White Hat facts):
[your bullet points from Step 2]

YOUR ROLE: You are the [COLOR] Hat in a Six Thinking Hats discussion (Round 3 of 3 — Final Round).
[same hat-specific persona as Round 1]

## Round 2 Reactions (all hats):

### Red Hat said:
[paste Red's Round 2 output]

### Black Hat said:
[paste Black's Round 2 output]

### Yellow Hat said:
[paste Yellow's Round 2 output]

### Green Hat said:
[paste Green's Round 2 output]

### Blue Hat said:
[paste Blue's Round 2 output]

ROUND 3 TASK: This is the final round. Deliver your concluding position:
- What is your hat's definitive recommendation?
- What one thing MUST happen based on this discussion?
- What was the most important insight that emerged from the debate?

RULES:
- Stay in your hat's perspective.
- Be concise and decisive — this is your final word.
- Keep your response to 100-150 words.
- End with "FINAL VERDICT:" — your single most important recommendation in one sentence.

Step 6: Synthesize

After all 3 rounds complete, write the final synthesis. Structure it as follows:

# Six Hats Analysis: [Topic]

**Date:** [current date]
**Rounds:** 3

---

## White Hat — Facts & Context

[Your bullet points from Step 2]

---

## Discussion Summary

### Round 1 — Opening Positions

#### Red Hat — Feelings & Intuition

[3-4 key bullets from Round 1]

#### Black Hat — Risks & Problems

[3-4 key bullets from Round 1]

#### Yellow Hat — Benefits & Value

[3-4 key bullets from Round 1]

#### Green Hat — Creative Ideas

[3-4 key bullets from Round 1]

#### Blue Hat — Process & Next Steps

[3-4 key bullets from Round 1]

### Round 2 — Key Debates

Summarize the most interesting exchanges between hats. Focus on:

- Where hats challenged each other
- Where positions shifted
- Where creative tension produced new insights

Write this as a narrative (2-3 paragraphs), not bullet points. Name the hats as speakers.

### Round 3 — Final Verdicts

| Hat | Final Verdict |
|-----|---------------|
| Red | [one sentence] |
| Black | [one sentence] |
| Yellow | [one sentence] |
| Green | [one sentence] |
| Blue | [one sentence] |

---

## Synthesis

[2-3 paragraphs weaving together the strongest insights from all 3 rounds.
Highlight how positions evolved through debate.
Identify the consensus that emerged AND the unresolved tensions.
End with a clear recommendation or set of options for the user.]

## Recommended Actions

[3-5 concrete, prioritized next steps that emerged from the discussion]

Step 7: Save results

If the user specified an output location, save the synthesis there. Otherwise, present it in the conversation.

Notes

• All 5 agents in each round MUST be launched in a single message for parallel execution.
• Rounds are sequential — Round 2 depends on Round 1, Round 3 depends on Round 2.
• Brief the user between rounds: "Round 1 complete, launching Round 2..." so they know progress is happening.
• Do NOT summarize what you're about to do at the start — just do it. The user knows what six hats means.
• If the topic is about the Klang project or codebase, ground your White Hat facts in actual code/architecture.
• The Round 2 debate summary is the most valuable part — this is where hats challenge each other and ideas evolve. Spend effort making it insightful, not just a recap.
• Keep agent prompts focused. Include only the previous round's outputs, not the full history of all rounds (to avoid context bloat).