event-storming

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Event Storming

Interactive Workshop Configuration

交互式工作坊配置

Use AskUserQuestion to configure the event storming session:

yaml

undefined

使用AskUserQuestion来配置Event Storming会话：

yaml

undefined

Question 1: Workshop Mode (MCP: Event Storming methodology)

question: "Which event storming mode do you need?" header: "Mode" options:

label: "Full Simulation (Recommended)" description: "All 5 personas debate in parallel, 6 phases (~15K tokens)"
label: "Quick" description: "Single analysis pass, no personas (~3K tokens)"
label: "Guided" description: "Interactive with user, personas on-demand"
label: "Codebase Analysis" description: "Discover events from existing code structure"

question: "Which event storming mode do you need?" header: "Mode" options:

label: "Full Simulation (Recommended)" description: "All 5 personas debate in parallel, 6 phases (~15K tokens)"
label: "Quick" description: "Single analysis pass, no personas (~3K tokens)"
label: "Guided" description: "Interactive with user, personas on-demand"
label: "Codebase Analysis" description: "Discover events from existing code structure"

Question 2: Domain Scope (MCP: DDD bounded context patterns)

question: "How complex is the domain you're exploring?" header: "Scope" options:

label: "Single Bounded Context" description: "One domain area, focused discovery"
label: "Multiple Contexts" description: "Identify 2-5 bounded context boundaries"
label: "Enterprise Domain" description: "Comprehensive cross-domain mapping"
label: "Unknown" description: "Help me determine scope first"


Use these responses to select the appropriate workshop mode and calibrate depth of analysis.

question: "How complex is the domain you're exploring?" header: "Scope" options:

label: "Single Bounded Context" description: "One domain area, focused discovery"
label: "Multiple Contexts" description: "Identify 2-5 bounded context boundaries"
label: "Enterprise Domain" description: "Comprehensive cross-domain mapping"
label: "Unknown" description: "Help me determine scope first"


根据这些回复选择合适的工作坊模式并调整分析深度。

When to Use This Skill

何时使用该技能

Use this skill when you need to:

Discover domain events for a business process
Identify commands and actors in a system
Find bounded context boundaries
Simulate a multi-stakeholder event storming workshop
Prepare for actual event storming facilitation
Analyze an existing codebase for domain events

Keywords: event storming, domain events, commands, actors, bounded contexts, aggregates, domain discovery, workshop simulation, multi-persona, DDD

在以下场景中使用本技能：

为业务流程发现领域事件
识别系统中的命令与参与者
寻找限界上下文边界
模拟多利益相关者的Event Storming工作坊
为实际Event Storming引导做准备
分析现有代码库以发现领域事件

关键词： event storming, domain events, commands, actors, bounded contexts, aggregates, domain discovery, workshop simulation, multi-persona, DDD

What is Event Storming?

什么是Event Storming？

Event Storming is a workshop-based method for collaborative domain discovery. Participants use sticky notes to map out:

Events (Orange) - Things that happen in the domain
Commands (Blue) - Actions that trigger events
Actors (Yellow) - Who issues commands
Aggregates (Yellow) - Business entities that handle commands
Read Models (Green) - Information needed for decisions
Policies (Purple) - Business rules and reactions
External Systems (Pink) - Outside integrations
Hot Spots (Red/Pink) - Areas of confusion or conflict

Event Storming是一种基于工作坊的协作式领域发现方法。参与者使用便利贴绘制以下内容：

事件（橙色）- 领域中发生的事情
命令（蓝色）- 触发事件的动作
参与者（黄色）- 发出命令的角色
聚合根（黄色大便利贴）- 处理命令的业务实体
查询模型（绿色）- 决策所需的信息
策略（紫色）- 业务规则与响应逻辑
外部系统（粉色）- 外部集成服务
热点（红/粉色）- 存在困惑或冲突的区域

Workshop Modes

工作坊模式

This skill supports three modes of operation:

Mode	Description	Token Cost	Use Case
`full-simulation`	All 5 personas debate in parallel, 6 phases	~15K tokens	Comprehensive discovery
`quick`	Single analysis pass, no personas	~3K tokens	Quick domain overview
`guided`	Interactive with user, personas on-demand	Variable	User wants control

本技能支持三种操作模式：

模式	描述	Token成本	使用场景
`full-simulation`	5个角色并行讨论，共6个阶段	~15000 tokens	全面领域发现
`quick`	单次分析，无角色模拟	~3000 tokens	快速领域概览
`guided`	与用户交互，按需调用角色	可变	用户希望掌控流程

Mode Selection Guide

模式选择指南

Use
full-simulation
when:

Starting a new project or major feature
You need comprehensive domain discovery
Multiple perspectives are valuable
Time is less critical than thoroughness

Use
quick
when:

You need a fast domain overview
Token budget is constrained
Domain is relatively simple
You'll refine later

Use
guided
when:

You want to drive the process
You have specific questions
You want to invoke specific personas
Interactive exploration is preferred

选择
full-simulation
当：

启动新项目或重大功能
需要全面的领域发现
多视角分析很有价值
时间充裕，更看重分析彻底性

选择
quick
当：

需要快速的领域概览
Token预算有限
领域相对简单
后续会进行细化分析

选择
guided
当：

希望主导流程
有特定问题需要解答
希望调用特定角色
偏好交互式探索

Multi-Persona Simulation

多角色模拟

The full-simulation mode uses 5 specialized agents to simulate different stakeholder perspectives:

Persona Agent	Role	Perspective	Contributions
`domain-expert`	Subject Matter Expert	Deep business knowledge	Domain events, business rules, edge cases
`developer-persona`	Technical Implementation	System constraints	Technical events, integration points
`business-analyst`	Process & Requirements	Process flow	Commands, actors, acceptance criteria
`product-owner`	Product Vision	User value	Priorities, MVP scope, user stories
`devils-advocate`	Challenger	Identify gaps	Hot spots, missing scenarios, contradictions

全模拟模式使用5个专业Agent来模拟不同利益相关者的视角：

角色Agent	角色	视角	贡献
`domain-expert`	领域专家	深入的业务知识	领域事件、业务规则、边缘场景
`developer-persona`	技术实现人员	系统约束	技术事件、集成点
`business-analyst`	流程与需求分析师	业务流程	命令、参与者、验收标准
`product-owner`	产品负责人	用户价值	优先级、MVP范围、用户故事
`devils-advocate`	挑刺者	发现漏洞	热点、缺失场景、矛盾点

6 Workshop Phases (Full Simulation)

6个工作坊阶段（全模拟）

Phase 1: Chaotic Exploration

阶段1：混沌探索

All personas brainstorm events independently. No constraints, no ordering.

Orchestration:

markdown

Launch 5 parallel Task agents:
- Task(event-storming-persona domain-expert, "Identify all domain events for: {domain}")
- Task(event-storming-persona developer, "Identify technical events for: {domain}")
- Task(event-storming-persona business-analyst, "Identify commands and actors for: {domain}")
- Task(event-storming-persona product-owner, "Prioritize and identify MVP scope for: {domain}")
- Task(event-storming-persona devils-advocate, "Challenge and identify gaps for: {domain}")

所有角色独立头脑风暴事件，无约束、无顺序要求。

编排逻辑：

markdown

Launch 5 parallel Task agents:
- Task(event-storming-persona domain-expert, "Identify all domain events for: {domain}")
- Task(event-storming-persona developer, "Identify technical events for: {domain}")
- Task(event-storming-persona business-analyst, "Identify commands and actors for: {domain}")
- Task(event-storming-persona product-owner, "Prioritize and identify MVP scope for: {domain}")
- Task(event-storming-persona devils-advocate, "Challenge and identify gaps for: {domain}")

Phase 2: Timeline Ordering

阶段2：时间线排序

Synthesize and order events chronologically. Create the timeline.

综合并按时间顺序排列事件，创建时间线。

Phase 3: Command Discovery

阶段3：命令发现

Identify what triggers each event. Map commands to events.

识别每个事件的触发因素，将命令与事件关联。

Phase 4: Actor Identification

阶段4：参与者识别

Map who issues each command. Define roles and systems.

映射发出每个命令的角色，定义角色与系统。

Phase 5: Bounded Context Discovery

阶段5：限界上下文发现

Group related events/commands. Identify natural boundaries.

对相关事件/命令进行分组，识别自然边界。

Phase 6: Hot Spot Resolution

阶段6：热点解决

Devil's advocate challenges. Resolve conflicts and gaps.

Detailed phase guidance: See

references/workshop-facilitation.md

挑刺者提出质疑，解决冲突与漏洞。

详细阶段指南： 参见

references/workshop-facilitation.md

Sticky Note Color Convention

便利贴颜色规范

Color	Represents	Example
🟧 Orange	Domain Event	"Order Placed", "Payment Received"
🟦 Blue	Command	"Place Order", "Process Payment"
🟨 Yellow (small)	Actor	"Customer", "Admin", "System"
🟨 Yellow (large)	Aggregate	"Order", "Customer", "Product"
🟩 Green	Read Model	"Order Summary", "Product Catalog"
🟪 Purple	Policy	"When order placed, reserve inventory"
🟫 Pink	External System	"Payment Gateway", "Email Service"
❗ Red/Pink	Hot Spot	Areas of confusion or conflict

Detailed conventions: See

references/sticky-note-types.md

颜色	代表内容	示例
🟧 橙色	领域事件	"订单已提交"、"付款已收到"
🟦 蓝色	命令	"提交订单"、"处理付款"
🟨 黄色（小）	参与者	"客户"、"管理员"、"系统"
🟨 黄色（大）	聚合根	"订单"、"客户"、"产品"
🟩 绿色	查询模型	"订单摘要"、"产品目录"
🟪 紫色	策略	"当订单提交时，预留库存"
🟫 粉色	外部系统	"支付网关"、"邮件服务"
❗ 红/粉色	热点	存在困惑或冲突的区域

详细规范： 参见

references/sticky-note-types.md

Orchestration Pattern

编排模式

Since Claude Code subagents cannot spawn other subagents, the main conversation orchestrates:

text

Main Conversation
    ↓
Invokes event-storming skill
    ↓
Skill guides parallel Task tool calls:
    ├── Task(event-storming-persona domain-expert, prompt)
    ├── Task(event-storming-persona developer, prompt)
    ├── Task(event-storming-persona business-analyst, prompt)
    ├── Task(event-storming-persona product-owner, prompt)
    └── Task(event-storming-persona devils-advocate, prompt)
    ↓
Skill synthesizes results with provenance tracking
    ↓
Outputs event catalog with [persona] attribution

由于Claude Code子Agent无法生成其他子Agent，由主对话负责编排：

text

Main Conversation
    ↓
Invokes event-storming skill
    ↓
Skill guides parallel Task tool calls:
    ├── Task(event-storming-persona domain-expert, prompt)
    ├── Task(event-storming-persona developer, prompt)
    ├── Task(event-storming-persona business-analyst, prompt)
    ├── Task(event-storming-persona product-owner, prompt)
    └── Task(event-storming-persona devils-advocate, prompt)
    ↓
Skill synthesizes results with provenance tracking
    ↓
Outputs event catalog with [persona] attribution

Quick Start

快速开始

Full Simulation Mode

全模拟模式

markdown

I want to run a full event storming simulation for an e-commerce order management system.

Please:
1. Launch all 5 persona agents in parallel
2. Have them analyze the domain
3. Synthesize their findings
4. Identify bounded contexts
5. Resolve any hot spots

markdown

I want to run a full event storming simulation for an e-commerce order management system.

Please:
1. Launch all 5 persona agents in parallel
2. Have them analyze the domain
3. Synthesize their findings
4. Identify bounded contexts
5. Resolve any hot spots

Quick Mode

快速模式

markdown

Give me a quick event storm overview for a subscription billing system.
Focus on the core happy path events.

markdown

Give me a quick event storm overview for a subscription billing system.
Focus on the core happy path events.

Guided Mode

引导模式

markdown

Let's do a guided event storming session for a hospital appointment system.
Start with the patient booking journey and I'll guide from there.

markdown

Let's do a guided event storming session for a hospital appointment system.
Start with the patient booking journey and I'll guide from there.

Output Format

输出格式

The event storming session produces a structured event catalog:

markdown

undefined

Event Storming会话会生成结构化的事件目录：

markdown

undefined

Event Storm: [Domain Name]

Event Catalog

[Bounded Context Name]

Events:

[Event Name] [Domain Expert] - [Description]
[Event Name] [Developer] - [Description]

Commands:

[Command Name] → [Event Name] [Business Analyst]

Actors:

[Actor Name]: [Commands they can issue]

Aggregates:

[Aggregate Name]: [Events it produces]

Policies:

[Policy Name]: [Trigger] → [Action]

Events:

[Event Name] [Domain Expert] - [Description]
[Event Name] [Developer] - [Description]

Commands:

[Command Name] → [Event Name] [Business Analyst]

Actors:

[Actor Name]: [Commands they can issue]

Aggregates:

[Aggregate Name]: [Events it produces]

Policies:

[Policy Name]: [Trigger] → [Action]

Bounded Contexts Identified

[Context Name]
- Core Domain / Supporting / Generic
- [Events in this context]

[Context Name]
- Core Domain / Supporting / Generic
- [Events in this context]

Hot Spots

[Issue] - [Resolution or TODO]


**Full template:** See `references/templates/event-storm-output.md`

[Issue] - [Resolution or TODO]


**完整模板：** 参见`references/templates/event-storm-output.md`

Integration with Other Skills

与其他技能的集成

Event storming connects with:

domain-storytelling - Stories feed into event discovery
modular-architecture - Bounded contexts become modules
fitness-functions - Module isolation tests
adr-management - Document bounded context decisions

Workflow:

text

Domain Storytelling → Event Storming → Modular Architecture
(understand "what")   (design "how")   (implement "where")

Event Storming可与以下技能集成：

domain-storytelling - 领域故事为事件发现提供输入
modular-architecture - 限界上下文可作为模块基础
fitness-functions - 模块隔离测试
adr-management - 记录限界上下文决策

工作流：

text

Domain Storytelling → Event Storming → Modular Architecture
(理解“是什么”)   (设计“怎么做”)   (实现“在哪里”)

Best Practices

最佳实践

Start with events - Events are facts, commands are debatable
Use past tense - "Order Placed" not "Place Order"
Explore boundaries - Events help find module boundaries
Embrace chaos - Initial brainstorming should be messy
Track provenance - Know which perspective each insight came from
Resolve hot spots - Don't leave conflicts unaddressed

从事件开始 - 事件是事实，命令存在争议
使用过去式 - 如“订单已提交”而非“提交订单”
探索边界 - 事件有助于发现模块边界
接纳混沌 - 初始头脑风暴应保持开放
跟踪来源 - 记录每个见解的视角来源
解决热点 - 不要遗留冲突与漏洞

References

参考资料

```
references/workshop-facilitation.md
```
- Phase orchestration and timing
```
references/persona-prompts.md
```
- Prompt templates for each persona
```
references/sticky-note-types.md
```
- Color conventions and usage
```
references/bounded-context-discovery.md
```
- Context identification patterns

references/templates/event-storm-output.md

- Output format with provenance

```
references/workshop-facilitation.md
```
- 阶段编排与时间规划
```
references/persona-prompts.md
```
- 各角色的提示词模板
```
references/sticky-note-types.md
```
- 便利贴类型与规范
```
references/bounded-context-discovery.md
```
- 上下文识别模式

references/templates/event-storm-output.md

- 带来源跟踪的输出模板

Version History

版本历史

v1.0.0 (2025-12-22): Initial release
- Multi-persona simulation with 5 agents
- Three workshop modes (full, quick, guided)
- 6-phase workshop structure
- Provenance tracking for insights
- Integration with domain-storytelling and modular-architecture

v1.0.0 (2025-12-22): 初始版本
- 支持5个Agent的多角色模拟
- 三种工作坊模式（全模拟、快速、引导式）
- 6阶段工作坊结构
- 见解来源跟踪
- 与domain-storytelling和modular-architecture集成

Last Updated

最后更新

Date: 2025-12-22 Model: claude-opus-4-5-20251101

日期： 2025-12-22 模型： claude-opus-4-5-20251101