brainstorm（头脑风暴）

Brainstorm

Facilitate structured ideation to explore options before commitment.

When to Use

• After vision is established, before PRD
• When multiple approaches seem viable
• When creative exploration is needed
• When stakeholders need to align on direction

Dependencies

• vision.md - Vision document must exist

- Check: projects/{project}/docs/vision.md

Book Triggers

Reference: [[books/thinking-rationality.md]]

• Apply "Scout Mindset" to explore genuinely, not confirm
• Use divergent-then-convergent thinking

Reference: [[books/product-innovation.md]]

• Apply "Inspired" discovery techniques

Brainstorming Protocol

Phase 1: Divergent (Generate)

Goal: Maximize option generation without judgment.

1. Set the frame

- Reference the vision document - State the specific question being explored - Set time limit for generation phase

1. Generate options

- Quantity over quality initially - No criticism of ideas - Build on others' ideas - Include "crazy" ideas

1. Capture everything

- List all ideas without filtering - Note who suggested what (if multi-agent)

Phase 2: Convergent (Evaluate)

Goal: Narrow to viable options.

1. Group similar ideas

- Cluster related approaches - Name each cluster

1. Initial filter

- Does it address the vision's problem? - Is it feasible given constraints? - Remove obviously non-viable options

1. Deep evaluation

- Apply decision-making skill - Consider tradeoffs - Assess against success criteria

Phase 3: Synthesis

Goal: Arrive at recommendation.

1. Shortlist (2-4 options)
2. Compare systematically
3. Recommend with rationale
4. Document dissent

Brainstorm Output Template

# Brainstorm: [Topic]

## TLDR

[2-3 sentences: What we explored, what emerged as the leading direction, and why.]

## Context

**Vision:** [Link to vision doc]
**Question:** [Specific question being explored]
**Participants:** [Who participated]

## Ideas Generated

### Category A: [Name]
- Idea 1
- Idea 2
- Idea 3

### Category B: [Name]
- Idea 4
- Idea 5

### Category C: [Name]
- Idea 6
- Idea 7

## Evaluation

### Shortlisted Options

#### Option 1: [Name]
- **Description:** [What it is]
- **Pros:** [Benefits]
- **Cons:** [Drawbacks]
- **Fit with vision:** [How well does it address the problem?]

#### Option 2: [Name]
- **Description:** [What it is]
- **Pros:** [Benefits]
- **Cons:** [Drawbacks]
- **Fit with vision:** [How well does it address the problem?]

### Comparison Matrix

| Criterion | Option 1 | Option 2 |
|-----------|----------|----------|
| Addresses problem | | |
| Feasibility | | |
| Time to value | | |
| Risk | | |

## Recommendation

**Recommended:** [Option name]
**Rationale:** [Why this option]

## Dissent

[Any disagreements or concerns about the recommendation]

## Next Steps

- [ ] Create PRD for recommended option
- [ ] [Other follow-up actions]

Output Location

projects/{project}/docs/brainstorm.md

Or for specific feature:

projects/{project}/docs/brainstorm-{feature}.md

Multi-Agent Brainstorming

When involving multiple agents:

1. Use facilitator skill

- Coordinate the session - Ensure all perspectives heard

1. Assign perspectives

- Technical feasibility (CTO/Architect) - User value (Product) - Business impact (CMO/CEO) - Risk (Conservative board member)

1. Structured rounds

- Round 1: Generate from your perspective - Round 2: React to others' ideas - Round 3: Synthesize and recommend

Quality Checklist

• Vision document referenced
• Multiple categories of ideas generated
• At least 2 options shortlisted
• Tradeoffs explicitly compared
• Recommendation has clear rationale
• Dissent documented if any

Anti-Patterns

• Premature convergence - Evaluating too early kills creativity
• Single option - Always explore alternatives
• Confirmation bias - Genuinely consider all options
• Ignoring constraints - Ideas must be feasible
• Missing dissent - Capture disagreements for record