AgentSkillsCN

brainstorming

通过结构化的设计头脑风暴，辅以清晰的提问与多视角的深度探索。当您需要新增功能、打造全新产品、实施各类集成，或在需求尚不明确时，可调用此技能。当用户提出“添加”“构建”“实施”“创建”“设计”等关键词，或发出开放式需求时，此技能便会自动触发。

SKILL.md
--- frontmatter
name: brainstorming
description: Conducts structured design brainstorming with clarifying questions and multi-perspective exploration. Use when adding features, building something new, implementing integrations, or when requirements are vague. Triggers on "add", "build", "implement", "create", "design", or open-ended requests.

Brainstorming

Structured Q&A sessions with parallel perspective exploration to clarify design goals before implementation.

When to Activate

  • •User mentions adding a new feature or integration
  • •User asks to "build", "implement", "create", or "add" something
  • •Requirements are vague or open-ended
  • •Multiple implementation approaches are possible
  • •Starting a new project or major feature

Process

1. Understand the Request

Parse the user's request to identify:

  • •Core objective - What they want to achieve
  • •Constraints - Time, tech stack, compatibility requirements
  • •Context - Existing codebase, dependencies, team preferences

2. Parallel Perspective Exploration

Spawn 2-3 agents to explore different angles simultaneously:

Technical Feasibility Agent:

  • •Analyze implementation complexity
  • •Identify technical constraints
  • •Evaluate tech stack fit
  • •Assess integration requirements

User Experience Agent:

  • •Explore user workflows
  • •Identify pain points addressed
  • •Consider edge cases in usage
  • •Map user journey

Risk Assessment Agent:

  • •Identify potential blockers
  • •Evaluate dependencies
  • •Assess scope creep risk
  • •Note security considerations

3. Explore the Codebase

Before asking questions, gather context:

  • •Search for related existing code
  • •Identify patterns and conventions used
  • •Note dependencies and integrations
  • •Find similar implementations for reference

4. Ask Clarifying Questions

Use AskUserQuestion tool for structured input. Focus on:

Scope Questions:

  • •What is the minimum viable version?
  • •What's must-have vs nice-to-have?
  • •What's explicitly out of scope?

Technical Questions:

  • •Preferred technologies or patterns?
  • •Integration points with existing systems?
  • •Performance or scalability requirements?
  • •Security/compliance requirements?

User Experience Questions:

  • •Who are the primary users?
  • •What workflows should this support?
  • •How should errors be handled?

5. Synthesize Design Direction

After gathering answers:

  1. •Summarize understood requirements
  2. •Aggregate insights from parallel explorations
  3. •Identify conflicts requiring resolution
  4. •Propose approach options with trade-offs
  5. •Recommend path forward with rationale

Human-in-the-Loop Rules

  • •If acceptance criteria are incomplete → Ask numbered questions and WAIT
  • •If conflicting requirements detected → Present trade-offs and WAIT
  • •If scope is ambiguous → Clarify boundaries before proceeding
  • •Never assume business decisions → Always confirm with user

Output Format

markdown
## Design Summary: [Feature Name]

### Objectives
- **Primary goal:** [Clear outcome]
- **Success criteria:** [Measurable indicators]
- **Out of scope:** [Explicit exclusions]

### Requirements Gathered
| Category | Requirement | Priority | Notes |
|----------|-------------|----------|-------|
| [Area] | [Req] | Must/Should/Could | [Context] |

### Technical Approach
- **Architecture:** [High-level design]
- **Key components:** [Major pieces]
- **Integration points:** [External systems]
- **Patterns to follow:** [Existing conventions]

### Trade-offs Considered
| Option | Pros | Cons | Recommendation |
|--------|------|------|----------------|
| [A] | [Benefits] | [Drawbacks] | [Yes/No] |

### Open Questions
- [Any remaining uncertainties]

### Risks Identified
- [Technical risks]
- [Scope risks]
- [Dependency risks]

### Recommended Next Step
[Clear action - typically invoke planning skill]

Agent Dispatch Templates

Technical Feasibility Agent

code
Analyze technical feasibility for:

## Feature
[Description]

## Current Stack
[Technologies in use]

## Tasks
1. Assess implementation complexity (S/M/L)
2. Identify required changes to existing code
3. Note integration challenges
4. Recommend approach

Be specific about technical trade-offs.

Risk Assessment Agent

code
Assess risks for proposed feature:

## Feature
[Description]

## Context
[Project state, timeline, dependencies]

## Tasks
1. Identify technical risks
2. Note scope creep potential
3. Assess dependency risks
4. Flag security considerations

Rank risks by impact and likelihood.

Collaboration

After brainstorming is complete, suggest invoking the planning skill to create a structured implementation plan with tasks and verification steps.