brainstorm

name

brainstorm

description

Interactive brainstorming and idea evaluation framework for generating and assessing novel ideas across any domain. Use when users want to: (1) Generate new ideas for projects, features, products, or strategies, (2) Brainstorm solutions to problems, (3) Evaluate and prioritize existing ideas, (4) Explore opportunities in software design, AI/ML, architecture, startups, or business planning, (5) Run structured ideation sessions, or (6) Analyze and compare multiple concepts. Triggers on requests for brainstorming, ideation, creative thinking, idea generation, concept evaluation, or opportunity exploration.

Interactive Brainstorming Framework

Conduct structured brainstorming sessions using proven techniques to generate, explore, and evaluate ideas across any domain.

Session Workflow

Every brainstorming session follows this four-phase structure:

Phase 1: Context Gathering

Start by understanding the scope and constraints through brief questions:

• Domain: What area are we exploring? (e.g., AI product, software architecture, business strategy)
• Objective: What outcome do we want? (e.g., solve a problem, explore opportunities, improve existing system)
• Constraints: Any key limitations? (e.g., budget, time, technical stack, team size)
• Background: Relevant context or prior attempts?

Keep this conversational and concise—aim for 3-5 questions maximum.

Phase 2: Divergent Ideation

Generate diverse ideas without immediate judgment using techniques from references/brainstorming-techniques.md:

Primary techniques:

• SCAMPER - Transform existing concepts (Substitute, Combine, Adapt, Modify, Put to other use, Eliminate, Reverse)
• First Principles - Break down to fundamentals and rebuild
• 5 Whys - Dig deeper to find root opportunities
• Analogical Thinking - Borrow solutions from other domains
• Constraint-Based - Force creativity through limitations

Approach:

1. Generate 5-10 ideas rapidly without filtering
2. For each idea, ask 1-2 follow-up questions to explore variations
3. Encourage wild ideas—evaluation comes later
4. Build on previous ideas when natural connections emerge

Output format:

Idea N: [Title]
- Core concept: [1-2 sentence description]
- Why this could work: [Key insight]
- Variation: [Alternative approach if relevant]

Phase 3: Convergent Evaluation

Assess each generated idea using quick feasibility-impact scoring:

Scoring dimensions (1-10 scale):

Feasibility - How achievable is this?

• Consider: Technical capability, resource availability, time required, dependencies
• 1-3: Requires breakthroughs | 4-6: Significant effort | 7-9: Achievable | 10: Trivial

Impact - How valuable is this?

• Consider: Business value, user benefit, strategic positioning, scale potential
• 1-3: Marginal | 4-6: Moderate | 7-9: Significant | 10: Transformational

Priority calculation:

Priority = (Feasibility + Impact) / 2
High: ≥7.5 | Medium: 5.0-7.4 | Low: <5.0

For each idea, provide:

• Feasibility score with brief rationale
• Impact score with brief rationale
• Priority designation (High/Medium/Low)
• Top 2-3 strengths
• Top 2-3 challenges or risks

See references/evaluation-frameworks.md for detailed assessment criteria.

Phase 4: Synthesis and Next Steps

Summarize the session and provide actionable guidance:

Summary format:

• Total ideas generated
• Priority distribution (High: N, Medium: N, Low: N)
• Top 2-3 recommendations with rationale
• Key insights or patterns observed

For high-priority ideas:

• Suggest 2-3 concrete next steps to validate or progress
• Identify critical unknowns or assumptions to test
• Recommend experiment approach if uncertainty is high

Save session: After completing the session, automatically save results using the save script:

# Generate JSON with session data
cat > /tmp/brainstorm_session.json << 'EOF'
{
  "session_title": "[Session Title]",
  "date": "[YYYY-MM-DD]",
  "domain": "[Domain]",
  "objective": "[Objective]",
  "context": "[Background context]",
  "ideas": [
    {
      "title": "[Idea Title]",
      "description": "[Description]",
      "category": "[Category/Type]",
      "feasibility": [1-10],
      "impact": [1-10],
      "priority": "[High/Medium/Low]",
      "strengths": ["strength 1", "strength 2", ...],
      "challenges": ["challenge 1", "challenge 2", ...],
      "next_steps": ["step 1", "step 2", ...]
    },
    ...
  ],
  "summary": "[Overall session summary]"
}
EOF

# Save to markdown
python3 scripts/save_ideas.py /tmp/brainstorm_session.json

The script saves to brainstorm-sessions/[timestamp].md by default.

Interactive Style

Conduct sessions in an interview format:

Tone:

• Conversational and energizing
• Ask thought-provoking questions
• Build on user responses
• Challenge assumptions constructively
• Celebrate creative leaps

Question patterns:

• "What if we...?"
• "How might we...?"
• "What would it look like if...?"
• "What's stopping us from...?"
• "How do [other domain] solve similar problems?"

Avoid:

• Premature criticism or dismissal
• Analysis paralysis
• Over-explaining methodologies
• Asking too many questions at once

Domain-Specific Adaptations

Tailor approach based on domain:

Software Architecture:

• Focus on trade-offs, scalability, maintainability
• Ask about non-functional requirements
• Consider future evolution and tech debt

AI/ML Projects:

• Evaluate data availability and quality
• Consider model complexity vs. value
• Assess operational challenges (monitoring, drift)

Startup Ideas:

• Emphasize market validation
• Explore go-to-market strategy
• Test business model assumptions

Product Features:

• Understand user jobs-to-be-done
• Consider adoption barriers
• Evaluate against product strategy

Business Strategy:

• Focus on competitive positioning
• Assess resource allocation
• Consider long-term defensibility

Analysis Tools

Analyze past sessions:

# Analyze single session
python3 scripts/analyze_ideas.py brainstorm-sessions/[filename].md

# Analyze all sessions
python3 scripts/analyze_ideas.py brainstorm-sessions/

# Generate report file
python3 scripts/analyze_ideas.py brainstorm-sessions/ -o report.md

Use analysis to identify patterns, track priorities, and surface recurring themes.

Templates

Quick idea capture: Use assets/idea-template.md for documenting individual ideas outside full sessions.

Session structure: Use assets/session-template.md as reference for manual documentation.

Advanced Techniques

When sessions need more structure, reference:

Brainstorming methods: See references/brainstorming-techniques.md for:

• Detailed SCAMPER examples
• First Principles breakdown process
• 5 Whys template
• Jobs to Be Done framework
• Reverse brainstorming
• Random input techniques
• Analogical thinking patterns

Evaluation frameworks: See references/evaluation-frameworks.md for:

• Detailed feasibility assessment criteria
• Extended impact dimensions
• Risk assessment frameworks
• ICE and RICE scoring models
• Strategic fit analysis
• Decision matrices

Load these references only when deeper methodology is needed.

Example Session Flow

User: "I need ideas for improving our API performance"

Claude:
Let's explore API performance improvements. Quick context:
1. What are your main performance bottlenecks? (e.g., database queries, network latency, computation)
2. What's your performance target? (e.g., <100ms response time, 10k RPS)
3. Any constraints? (e.g., can't change database, limited budget)

[User responds]

Great! Let me generate some ideas using different angles...

Idea 1: Response Caching with Smart Invalidation
- Core concept: Implement multi-layer caching (CDN + Redis + in-memory) with intelligent invalidation
- Why this could work: Dramatically reduces database load for read-heavy endpoints
- Variation: Use cache-aside pattern vs. write-through depending on consistency needs

[Continue with 4-9 more ideas using SCAMPER, First Principles, etc.]

Now let's evaluate these...

Idea 1: Response Caching
- Feasibility: 8/10 (Established patterns, requires cache infrastructure)
- Impact: 9/10 (Can reduce 70-90% of database load)
- Priority: HIGH
- Strengths: Proven approach, immediate ROI, scales horizontally
- Challenges: Cache invalidation complexity, memory costs, consistency trade-offs

[Continue evaluation for all ideas]

Summary:
Generated 7 performance improvement ideas. 3 high-priority (caching, query optimization, connection pooling), 3 medium-priority (compression, pagination, async processing), 1 low-priority (edge computing - high setup cost).

Top recommendations:
1. Start with response caching - highest impact/feasibility ratio
2. Parallel track: Optimize N+1 queries - foundational improvement
3. Later phase: Add read replicas if load continues growing

Next steps for caching approach:
1. Profile top 20 API endpoints by traffic
2. Implement Redis cache for top 5 endpoints as pilot
3. Measure impact and tune TTL/invalidation strategy

[Save session automatically]
✅ Session saved to: brainstorm-sessions/20260206-143022.md

Tips for Effective Sessions

Do:

• Start broad, then focus on promising directions
• Use specific examples to ground abstract ideas
• Challenge assumptions with "What if?" questions
• Connect ideas to show relationships
• Balance creativity with practicality

Don't:

• Rush to evaluation before generating enough ideas
• Dismiss unusual ideas too quickly
• Over-analyze during generation phase
• Ignore constraints entirely
• Make decisions without assessment