Agent Skills: Brainstorming

Brainstorming

Overview

This skill guides collaborative dialogue to transform ideas into detailed design specifications before implementation begins. Through Socratic questioning and iterative refinement, it ensures shared understanding and prevents costly rework.

Quick Start

1. Understand context - Examine project, ask clarifying questions
2. Explore options - Propose 2-3 approaches with trade-offs
3. Refine design - Validate section by section (200-300 words each)
4. Document - Write design to timestamped file
5. Plan - Optionally create implementation plan

When to Use

• Starting new features or projects
• Clarifying ambiguous requirements
• Evaluating architectural decisions
• Designing APIs or interfaces
• Planning complex implementations
• Before writing any significant code

The Brainstorming Process

Phase 1: Understanding

Goal: Build complete picture of the problem space.

Approach:

• Examine project context first
• Ask one question at a time
• Use multiple-choice questions when feasible
• Focus on purpose, constraints, and success metrics

Key questions:

• What problem are we solving?
• Who are the users?
• What are the constraints?
• How will success be measured?
• What already exists?

Phase 2: Exploration

Goal: Identify and evaluate solution approaches.

Approach:

• Propose 2-3 different approaches
• Present trade-offs for each
• Give reasoned recommendations
• Stay conversational, not prescriptive

Option template:

### Option A: [Name]

- Approach: [Description]
- Pros: [Benefits]
- Cons: [Drawbacks]
- Best for: [Scenarios]

Phase 3: Design Presentation

Goal: Create validated design specification.

Approach:

• Break into sections of 200-300 words
• Validate each section before proceeding
• Cover: architecture, components, data flow, error handling, testing
• Allow revisiting earlier decisions

Section checklist:

• [ ] Architecture overview

See sub-skills for full details.

Key Principles

YAGNI (You Aren't Gonna Need It)

Apply ruthlessly:

• Design only what's needed now
• Avoid speculative features
• Question every "nice to have"
• Defer complexity until required

Single Question Per Message

• Prevents overwhelming stakeholders
• Ensures each point is addressed
• Maintains conversation flow
• Allows for course correction

Incremental Validation

• Validate section by section
• Get explicit confirmation
• Allow reversals
• Build on confirmed foundations

Question Templates

Clarification

• "To clarify: [summary of understanding]. Is that correct?"
• "When you say [term], do you mean (a) [option1], (b) [option2], or (c) something else?"

Trade-off Exploration

• "We could either [A] or [B]. [A] gives us [benefit] but [drawback]. [B] gives us [benefit] but [drawback]. Which matters more for this project?"

Priority Assessment

• "Which is more important: [quality A] or [quality B]?"
• "If we had to choose between [option 1] and [option 2], which would you prefer?"

Validation

• "Here's my understanding of [section]. Does this match your expectations?"
• "Before we move on, let me confirm: [summary]"

Post-Design Actions

After validation:

1. Document

   • Write design to timestamped file
   • Include all decisions and rationale
   • Note any deferred decisions

2. Plan (Optional)

   • Use writing-plans skill for implementation
   • Break design into tasks
   • Estimate and prioritize

3. Review (Optional)

   • Share with stakeholders
   • Gather additional feedback
   • Incorporate changes

Related Skills

• writing-plans - Create implementation plans
• sparc-workflow - Development methodology
• product-roadmap - Product planning

Version History

• 1.0.0 (2026-01-19): Initial release adapted from obra/superpowers

Sub-Skills

• Best Practices
• Error Handling
• Metrics