brainstorming（头脑风暴）

Brainstorming Ideas Into Designs

Overview

Help turn ideas into fully formed designs and specs through natural collaborative dialogue.

Start by understanding the current project context, then ask questions one at a time to refine the idea. Once you understand what you're building, present the design in small sections (200-300 words), checking after each section whether it looks right so far.

When to Activate

Activate this skill when:

• User has a rough idea that needs refinement
• User is uncertain about scope or approach
• User says "I'm not sure what we need" or "I'm not sure how to approach this"
• During Phase 2 (Requirements) to explore WHAT to build
• During Phase 3 (Technical Design) to explore HOW to build it

Phase 2 Focus (Requirements):

• Clarifying what the feature should do
• Exploring different requirement scopes
• Understanding user needs and constraints
• Determining must-haves vs. nice-to-haves

Phase 3 Focus (Design):

• Exploring architectural approaches
• Comparing technical solutions
• Analyzing implementation trade-offs

Research-Driven Exploration

When exploring unclear requirements or designs, use web research to inform decisions:

For Requirements Exploration:

• Use WebSearch to find how others have solved similar problems
• Use WebFetch to analyze competitor features and documentation
• Research user expectations and industry standards

For Design Exploration:

• Use WebSearch to find architectural patterns and best practices
• Use WebFetch to read library/framework documentation
• Research performance benchmarks and case studies

For API Integration:

• Use Bash with curl to explore and test API endpoints
• Fetch API schemas and documentation
• Understand rate limits, authentication, and constraints

Research Questions to Answer:

• "How do similar products handle this?"
• "What are common pitfalls to avoid?"
• "What's the current best practice (2025)?"
• "Are there established patterns for this problem?"

🗣 Say: "Let me research this before proposing options. I'll look for prior art and best practices."

Document Research:

• Share key findings with user before presenting options
• Include sources for transparency
• Note any conflicting approaches found

The Process

Understanding the idea:

• Check out the current project state first (files, docs, recent commits)
• Ask questions one at a time to refine the idea
• Prefer multiple choice questions when possible, but open-ended is fine too
• Only one question per message - if a topic needs more exploration, break it into multiple questions
• Focus on understanding: purpose, constraints, success criteria

When to UltraThink: Before proposing architectural approaches, activate deep thinking if:

• Multiple valid solutions exist with significant trade-offs
• Decision has long-term architectural implications
• Requirements involve complex system interactions
• Security or performance are critical concerns

🗣 Say: "Let me ultrathink this before proposing approaches. I'll question fundamentals and consider implications from first principles."

During UltraThink:

• Question assumptions about requirements
• Consider second-order effects of each approach
• Think about what could go wrong (pre-mortem)
• Evaluate against similar systems you've seen
• Consider future maintainability and evolution

After UltraThink: Provide approaches with clear reasoning about trade-offs and long-term implications.

Exploring approaches:

• Propose 2-3 different approaches with trade-offs
• Present options conversationally with your recommendation and reasoning
• Lead with your recommended option and explain why
• Consider:

- Complexity (Low/Medium/High) - Maintainability - Performance implications - Security considerations - Testability

Presenting the design:

• Once you believe you understand what you're building, present the design
• Break it into sections of 200-300 words
• Ask after each section whether it looks right so far
• Cover: architecture, components, data flow, error handling, testing
• Be ready to go back and clarify if something doesn't make sense

After Brainstorming

For Requirements (Phase 2):

• Write validated requirements to docx/features/[NN-feature-name]/requirements.md
• Use EARS format (Event-Driven, State-Driven, Ubiquitous, Conditional, Optional)
• Include:

- Overview - Functional requirements - Non-functional requirements (performance, security, usability) - Constraints - Acceptance criteria - Out of scope items

• Ask: "Requirements complete. Ready for design phase?"

For Design (Phase 3):

• Write the validated design to docx/features/[NN-feature-name]/design.md
• Include:

- Architecture Overview - Component Structure - Data Flow - API Contracts - Error Handling Strategy - Security Considerations - Performance Considerations - Testing Strategy

• Ask: "Design complete. Ready for task breakdown?"

Transition:

• After requirements → Proceed to Phase 3 (Technical Design)
• After design → Transition to spec-driven-implementation skill for task breakdown

Key Principles

• One question at a time - Don't overwhelm with multiple questions
• Multiple choice preferred - Easier to answer than open-ended when possible
• YAGNI ruthlessly - Remove unnecessary features from all designs
• Explore alternatives - Always propose 2-3 approaches before settling
• Incremental validation - Present design in sections, validate each
• Be flexible - Go back and clarify when something doesn't make sense
• Think about testing - Good designs are testable designs
• Consider security - Build security in, don't bolt it on later

Example Questioning Flow

Understanding Purpose:

Q: "What's the primary goal of this authentication feature?"
→ User answers

Q: "Should it support multiple auth methods (email/password, OAuth, etc.)
   or just one method initially?"
→ User answers

Q: "What happens when a session expires - force re-login or offer refresh?"
→ User answers

Exploring Approaches:

Based on your requirements, I see 3 main approaches:

**Option A: JWT-Based Authentication** [RECOMMENDED]
Pros: Stateless, scalable, works across services, standard
Cons: Token invalidation complexity, larger payload
Complexity: Medium
Best for: Microservices, APIs, future scalability

**Option B: Session-Based Authentication**
Pros: Simple invalidation, smaller cookies, familiar
Cons: Requires session storage, scaling challenges
Complexity: Low
Best for: Monolithic apps, simple use cases

**Option C: Hybrid Approach**
Pros: Combines benefits of both
Cons: More complex, harder to maintain
Complexity: High
Best for: Complex enterprise requirements

I recommend Option A (JWT) because your requirements mention
potential API integrations and future mobile app support.
JWT is industry-standard for this use case.

Does this align with your thinking?

Presenting Design Incrementally:

Let me present the architecture in sections:

**Section 1: High-Level Flow**
[200-300 words describing auth flow]

Does this look right so far?
→ User validates or requests changes

**Section 2: Component Structure**
[200-300 words describing components]

How does this look?
→ Continue...

Integration with Spec-Driven Workflow

This skill can be used in two phases:

Phase 2 (Requirements):

• Use when user has rough idea but unclear requirements
• Helps clarify what to build vs. what's out of scope
• Explores different feature scopes and priorities
• Outputs to requirements.md in EARS format

Phase 3 (Technical Design):

• Use after requirements are defined
• Helps explore how to build the feature
• Compares architectural approaches with trade-offs
• Outputs to design.md with complete technical specs

After brainstorming completes:

• Phase 2 → Proceed to Phase 3 (Technical Design)
• Phase 3 → Transition to spec-driven-implementation for task breakdown

Notes

• Phase 2: Focus on "what" (what should system do? what's in/out of scope?)
• Phase 3: Focus on "how" (how should we build it? what are trade-offs?)
• Always present multiple options before recommending
• Validate incrementally - don't dump everything at once
• Be ready to backtrack if something doesn't make sense
• One question at a time - let user think and respond
• Good requirements lead to good designs
• Good designs enable good tests - think about testability
• Security and error handling are not afterthoughts