elicitation-methodology

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Elicitation Methodology

需求获取方法论

Central hub for requirements elicitation methodology, technique selection, and workflow orchestration.

需求获取方法论、技术选择和工作流编排的核心枢纽。

When to Use This Skill

何时使用此Skill

Keywords: requirements gathering, elicitation, stakeholder needs, requirement discovery, user needs, feature requests, interview, requirements session

Invoke this skill when:

Starting a new requirements elicitation effort
Selecting appropriate elicitation techniques
Orchestrating multi-source elicitation
Configuring autonomy levels for AI assistance
Understanding LLMREI interview patterns

关键词: 需求收集, 需求获取, 利益相关方需求, 需求挖掘, 用户需求, 功能请求, 访谈, 需求会议

在以下场景调用此Skill：

启动新的需求获取工作时
选择合适的需求获取技术时
编排多来源需求获取工作时
配置AI辅助的自主级别时
了解LLMREI访谈模式时

Quick Decision Tree

快速决策树

Scenario	Recommended Approach
Have stakeholders to interview	Use `interview-conducting` skill
Have documents/PDFs to mine	Use `document-extraction` skill
Working solo, need perspectives	Use `stakeholder-simulation` skill
Need domain knowledge	Use `domain-research` skill
Checking completeness	Use `gap-analysis` skill
Ready for specification	Use `/export` command

场景	推荐方案
有可访谈的利益相关方	使用 `interview-conducting` skill
有文档/PDF需要挖掘	使用 `document-extraction` skill
独立工作，需要多视角	使用 `stakeholder-simulation` skill
需要领域知识	使用 `domain-research` skill
检查需求完整性	使用 `gap-analysis` skill
准备好输出规格说明书	使用 `/export` 命令

Elicitation Techniques

需求获取技术

1. Stakeholder Interviews (LLMREI Pattern)

1. 利益相关方访谈（LLMREI模式）

AI-conducted interviews using research-backed prompting strategies.

When to use:

Direct access to stakeholders
Complex domains requiring exploration
Need to capture tacit knowledge

Technique reference: See

references/llmrei-patterns.md

采用基于研究的提示策略，由AI主导的访谈。

适用场景：

可直接接触利益相关方
复杂领域需要深入探索
需要捕捉隐性知识

技术参考： 查看

references/llmrei-patterns.md

2. Document Extraction

2. 文档提取

Mine requirements from existing documentation.

When to use:

Existing requirements documents
Meeting transcripts
Regulatory documents
Competitor analysis

Delegate to:

document-extraction

skill

从现有文档中挖掘需求。

适用场景：

已有需求文档
会议记录
监管文档
竞品分析报告

委托给：

document-extraction

skill

3. Stakeholder Simulation

3. 利益相关方模拟

Multi-persona simulation for solo requirements work.

When to use:

Working without direct stakeholder access
Need diverse perspectives
Validating completeness

Delegate to:

stakeholder-simulation

skill

为独立需求工作提供多角色模拟。

适用场景：

无法直接接触利益相关方
需要多样化视角
验证需求完整性

委托给：

stakeholder-simulation

skill

4. Domain Research

4. 领域研究

MCP-powered research for domain knowledge.

When to use:

Unfamiliar domain
Need industry standards
Competitive analysis
Technology constraints

Delegate to:

domain-research

skill

由MCP支持的领域研究。

适用场景：

不熟悉的领域
需要行业标准
竞品分析
技术约束调研

委托给：

domain-research

skill

Autonomy Levels

自主级别

Guided Mode (Human-in-Loop)

引导模式（人在回路）

yaml

autonomy: guided
behavior:
  - AI suggests questions, human approves
  - Each requirement validated individually
  - Human controls interview flow
  - Maximum transparency
use_when:
  - Sensitive or regulated domains
  - Learning the elicitation process
  - High-stakes requirements

yaml

autonomy: guided
behavior:
  - AI提出问题，由人工审批
  - 每条需求单独验证
  - 人工控制访谈流程
  - 最高透明度
use_when:
  - 敏感或受监管的领域
  - 学习需求获取流程
  - 高风险需求项目

Semi-Autonomous Mode

半自主模式

yaml

autonomy: semi-auto
behavior:
  - AI conducts interviews with checkpoints
  - Human validates requirement batches
  - Periodic progress reviews
  - Balance of speed and control
use_when:
  - Standard elicitation projects
  - Moderate domain complexity
  - Trusted AI capabilities

yaml

autonomy: semi-auto
behavior:
  - AI主导访谈，设置检查点
  - 人工批量验证需求
  - 定期进度回顾
  - 平衡速度与控制
use_when:
  - 标准需求获取项目
  - 中等复杂度领域
  - 信任AI能力的场景

Fully Autonomous Mode

全自主模式

yaml

autonomy: full-auto
behavior:
  - Complete end-to-end elicitation
  - Human reviews final output only
  - Maximum efficiency
  - AI handles all decisions
use_when:
  - Well-understood domains
  - Time pressure
  - Preliminary discovery

yaml

autonomy: full-auto
behavior:
  - 端到端完整需求获取
  - 人工仅审核最终输出
  - 最高效率
  - AI处理所有决策
use_when:
  - 熟悉的领域
  - 时间紧张
  - 初步需求挖掘

Workflow Orchestration

工作流编排

Standard Discovery Workflow

标准挖掘工作流

text

1. CONTEXT GATHERING
   ├── Load any existing business context
   ├── Identify available sources (stakeholders, docs, etc.)
   └── Select autonomy level

2. MULTI-SOURCE ELICITATION
   ├── Interviews (if stakeholders available)
   ├── Document extraction (if docs available)
   ├── Domain research (MCP queries)
   └── Stakeholder simulation (if solo mode)

3. SYNTHESIS
   ├── Consolidate requirements from all sources
   ├── Remove duplicates
   ├── Classify by type (functional, NFR, constraint)
   └── Apply MoSCoW prioritization

4. VALIDATION
   ├── Gap analysis
   ├── Completeness checking
   ├── Conflict detection
   └── INVEST scoring

5. OUTPUT
   ├── Save to .requirements/{domain}/
   ├── Generate summary report
   └── Prepare for specification export

text

1. 上下文收集
   ├── 加载现有业务上下文
   ├── 识别可用来源（利益相关方、文档等）
   └── 选择自主级别

2. 多来源需求获取
   ├── 访谈（如有利益相关方）
   ├── 文档提取（如有文档）
   ├── 领域研究（MCP查询）
   └── 利益相关方模拟（如独立工作）

3. 需求整合
   ├── 整合所有来源的需求
   ├── 去除重复项
   ├── 按类型分类（功能需求、非功能需求、约束）
   └── 应用MoSCoW优先级排序

4. 验证
   ├── 差距分析
   ├── 完整性检查
   ├── 冲突检测
   └── INVEST评分

5. 输出
   ├── 保存至.requirements/{domain}/
   ├── 生成总结报告
   └── 准备导出为规格说明书

Output Format

输出格式

Pre-Canonical Requirements

预规范需求

yaml

undefined

yaml

undefined

.requirements/{domain}/requirements.yaml

id: REQ-SET-{number} title: "{Domain} Requirements" domain: "{domain-name}" elicitation_date: "{ISO-8601-date}" autonomy_level: "{guided|semi-auto|full-auto}"

sources:

type: interview|document|simulation|research reference: "{source-identifier}" timestamp: "{ISO-8601-date}"

requirements:

id: REQ-{number} text: "{requirement statement}" source: "{source-type}" source_ref: "{specific-reference}" priority: must|should|could|wont category: functional|non-functional|constraint|assumption confidence: high|medium|low validation_status: pending|validated|rejected

gaps_identified:

category: "{requirement-category}" description: "{what's missing}" severity: critical|major|minor

metadata: total_sources: {number} total_requirements: {number} gap_count: {number} ready_for_specification: true|false

undefined

id: REQ-SET-{number} title: "{Domain} Requirements" domain: "{domain-name}" elicitation_date: "{ISO-8601-date}" autonomy_level: "{guided|semi-auto|full-auto}"

sources:

type: interview|document|simulation|research reference: "{source-identifier}" timestamp: "{ISO-8601-date}"

requirements:

id: REQ-{number} text: "{需求描述}" source: "{来源类型}" source_ref: "{具体参考}" priority: must|should|could|wont category: functional|non-functional|constraint|assumption confidence: high|medium|low validation_status: pending|validated|rejected

gaps_identified:

category: "{需求类别}" description: "{缺失内容}" severity: critical|major|minor

metadata: total_sources: {number} total_requirements: {number} gap_count: {number} ready_for_specification: true|false

undefined

Export Options

导出选项

After elicitation, requirements can be exported to various specification formats:

bash

/requirements-elicitation:export --to canonical  # Canonical spec format
/requirements-elicitation:export --to ears       # EARS pattern format
/requirements-elicitation:export --to gherkin    # Gherkin/BDD format

需求获取完成后，可将需求导出为多种规格说明书格式：

bash

/requirements-elicitation:export --to canonical  # 规范格式
/requirements-elicitation:export --to ears       # EARS模式格式
/requirements-elicitation:export --to gherkin    # Gherkin/BDD格式

Related Skills

References

参考资料

```
references/llmrei-patterns.md
```
- LLMREI prompting strategies
```
references/technique-matrix.md
```
- Technique selection guidance
```
references/autonomy-levels.md
```
- Detailed autonomy configuration

Last Updated: 2025-12-26

```
references/llmrei-patterns.md
```
- LLMREI提示策略
```
references/technique-matrix.md
```
- 技术选择指南
```
references/autonomy-levels.md
```
- 详细自主级别配置

最后更新: 2025-12-26

elicitation-methodology

Original

Translation

Elicitation Methodology

需求获取方法论

When to Use This Skill

何时使用此Skill

Quick Decision Tree

快速决策树

Elicitation Techniques

需求获取技术

1. Stakeholder Interviews (LLMREI Pattern)

1. 利益相关方访谈（LLMREI模式）

2. Document Extraction

2. 文档提取

3. Stakeholder Simulation

3. 利益相关方模拟

4. Domain Research

4. 领域研究

Autonomy Levels

自主级别

Guided Mode (Human-in-Loop)

引导模式（人在回路）

Semi-Autonomous Mode

半自主模式

Fully Autonomous Mode

全自主模式

Workflow Orchestration

工作流编排

Standard Discovery Workflow

标准挖掘工作流

Output Format

输出格式

Pre-Canonical Requirements

预规范需求

.requirements/{domain}/requirements.yaml

.requirements/{domain}/requirements.yaml

Export Options

导出选项

Related Skills

相关Skills

References

参考资料