pdd-skills 3.0.0

package/skills/core/official-doc-writer/scripts/smart_prompts.py

@@ -0,0 +1,363 @@
+# -*- coding: utf-8 -*-
+"""
+公文智能提示系统
+根据公文类型和上下文提供智能提示
+"""
+
+from typing import Dict, List, Optional
+from enum import Enum
+
+
+class DocumentType(Enum):
+    """公文类型枚举"""
+    NOTICE = "通知"
+    REPORT = "报告"
+    REQUEST = "请示"
+    LETTER = "函"
+    CIRCULAR = "通报"
+    MINUTES = "纪要"
+    DECISION = "决定"
+    OPINION = "意见"
+    REPLY = "批复"
+    ORDER = "命令(令)"
+
+
+class SmartPromptSystem:
+    """智能提示系统"""
+    
+    def __init__(self):
+        self.opening_prompts = self._init_opening_prompts()
+        self.ending_prompts = self._init_ending_prompts()
+        self.structure_hints = self._init_structure_hints()
+        self.language_templates = self._init_language_templates()
+    
+    def _init_opening_prompts(self) -> Dict[str, List[str]]:
+        """初始化开头语提示"""
+        return {
+            "通知": [
+                "为{subject}，现{action}如下：",
+                "根据{basis}，决定{action}。",
+                "为了{purpose}，现将{content}通知如下：",
+                "经{authority}研究决定，现就{subject}通知如下："
+            ],
+            "报告": [
+                "现将{subject}报告如下：",
+                "关于{subject}情况报告如下：",
+                "根据{requirement}，现将{content}报告如下：",
+                "按照{instruction}，现就{subject}报告如下："
+            ],
+            "请示": [
+                "关于{subject}的请示",
+                "现就{subject}请示如下：",
+                "为{purpose}，拟{action}，现请示如下："
+            ],
+            "函": [
+                "关于{subject}的函",
+                "现就{subject}函告如下：",
+                "为{purpose}，商请{action}，现函告如下："
+            ],
+            "通报": [
+                "现将{subject}通报如下：",
+                "关于{subject}情况的通报",
+                "为{purpose}，现将{content}通报如下："
+            ],
+            "纪要": [
+                "{meeting_name}纪要",
+                "{date}，{organizer}召开{meeting_name}，现将会议主要情况纪要如下："
+            ],
+            "决定": [
+                "为{purpose}，决定{action}。",
+                "经{authority}研究，决定{action}。"
+            ],
+            "意见": [
+                "为{purpose}，现提出如下意见：",
+                "关于{subject}的意见",
+                "根据{basis}，现就{subject}提出如下意见："
+            ],
+            "批复": [
+                "你{unit}《关于{subject}的请示》收悉。现批复如下：",
+                "关于{subject}的批复"
+            ],
+            "命令(令)": [
+                "为{purpose}，现发布命令如下：",
+                "根据{authority}，现发布{order_type}："
+            ]
+        }
+    
+    def _init_ending_prompts(self) -> Dict[str, List[str]]:
+        """初始化结尾语提示"""
+        return {
+            "通知": [
+                "特此通知。",
+                "请认真贯彻执行。",
+                "请遵照执行。",
+                "以上通知，请认真贯彻落实。"
+            ],
+            "报告": [
+                "特此报告。",
+                "以上报告，请审阅。",
+                "以上报告如有不妥，请指正。",
+                "专此报告。"
+            ],
+            "请示": [
+                "妥否，请批示。",
+                "以上请示，请予批复。",
+                "当否，请批示。",
+                "以上请示如无不妥，请批复。"
+            ],
+            "函": [
+                "请予研究函复。",
+                "特此函告。",
+                "盼复。",
+                "请予支持为盼。",
+                "特此函达。"
+            ],
+            "通报": [
+                "特此通报。",
+                "望各单位引以为戒。",
+                "希各单位认真学习借鉴。"
+            ],
+            "纪要": [
+                "（纪要一般不写结尾语）"
+            ],
+            "决定": [
+                "（决定一般不写结尾语）",
+                "本决定自发布之日起施行。"
+            ],
+            "意见": [
+                "以上意见，请结合实际贯彻执行。",
+                "以上意见如无不妥，请批转各地执行。"
+            ],
+            "批复": [
+                "此复。",
+                "特此批复。"
+            ],
+            "命令(令)": [
+                "（命令一般不写结尾语）",
+                "此令。"
+            ]
+        }
+    
+    def _init_structure_hints(self) -> Dict[str, List[str]]:
+        """初始化结构提示"""
+        return {
+            "通知": [
+                "一、{背景和目的}",
+                "二、{主要内容}",
+                "三、{具体要求}",
+                "四、{执行时间}"
+            ],
+            "报告": [
+                "一、{工作背景}",
+                "二、{主要做法}",
+                "三、{取得成效}",
+                "四、{存在问题}",
+                "五、{下一步打算}"
+            ],
+            "请示": [
+                "一、{请示事由}",
+                "二、{请示事项}",
+                "三、{请示理由}",
+                "四、{建议方案}"
+            ],
+            "函": [
+                "一、{发函事由}",
+                "二、{主要内容}",
+                "三、{请求事项}"
+            ],
+            "通报": [
+                "一、{通报事由}",
+                "二、{基本情况}",
+                "三、{经验教训}",
+                "四、{工作要求}"
+            ],
+            "纪要": [
+                "会议时间：{time}",
+                "会议地点：{location}",
+                "参会人员：{participants}",
+                "会议议题：{agenda}",
+                "会议内容：{content}",
+                "会议决定：{decisions}"
+            ],
+            "决定": [
+                "一、{决定事由}",
+                "二、{决定事项}",
+                "三、{执行要求}"
+            ],
+            "意见": [
+                "一、{指导思想}",
+                "二、{基本原则}",
+                "三、{主要目标}",
+                "四、{重点任务}",
+                "五、{保障措施}"
+            ],
+            "批复": [
+                "一、{批复依据}",
+                "二、{批复意见}",
+                "三、{执行要求}"
+            ],
+            "命令(令)": [
+                "一、{命令事由}",
+                "二、{命令事项}",
+                "三、{执行要求}"
+            ]
+        }
+    
+    def _init_language_templates(self) -> Dict[str, Dict[str, str]]:
+        """初始化语言模板"""
+        return {
+            "通知": {
+                "subject": "事由",
+                "action": "行动",
+                "basis": "依据",
+                "purpose": "目的",
+                "content": "内容",
+                "authority": "权威机关"
+            },
+            "报告": {
+                "subject": "报告事项",
+                "requirement": "要求",
+                "instruction": "指示",
+                "content": "内容"
+            },
+            "请示": {
+                "subject": "请示事项",
+                "purpose": "目的",
+                "action": "行动"
+            },
+            "函": {
+                "subject": "函告事项",
+                "purpose": "目的",
+                "action": "行动"
+            }
+        }
+    
+    def get_opening_prompt(self, doc_type: str, context: Optional[Dict] = None) -> List[str]:
+        """获取开头语提示"""
+        prompts = self.opening_prompts.get(doc_type, [])
+        
+        if context:
+            filled_prompts = []
+            for prompt in prompts:
+                try:
+                    filled_prompt = prompt.format(**context)
+                    filled_prompts.append(filled_prompt)
+                except KeyError:
+                    filled_prompts.append(prompt)
+            return filled_prompts
+        
+        return prompts
+    
+    def get_ending_prompt(self, doc_type: str) -> List[str]:
+        """获取结尾语提示"""
+        return self.ending_prompts.get(doc_type, [])
+    
+    def get_structure_hint(self, doc_type: str) -> List[str]:
+        """获取结构提示"""
+        return self.structure_hints.get(doc_type, [])
+    
+    def get_language_template(self, doc_type: str) -> Dict[str, str]:
+        """获取语言模板"""
+        return self.language_templates.get(doc_type, {})
+    
+    def suggest_title(self, doc_type: str, issuer: str, subject: str) -> str:
+        """建议标题"""
+        return f"{issuer}关于{subject}的{doc_type}"
+    
+    def suggest_doc_number(self, issuer_abbr: str, year: int, seq: int) -> str:
+        """建议发文字号"""
+        return f"{issuer_abbr}〔{year}〕{seq}号"
+    
+    def get_contextual_hint(self, doc_type: str, field: str) -> str:
+        """获取上下文提示"""
+        hints = {
+            "issuer": "发文机关是公文的制发机关，应使用全称或规范化简称",
+            "recipient": "主送机关是公文的主要受理机关，应当使用全称、规范化简称或同类型机关统称",
+            "title": "标题由发文机关名称、事由和文种组成，应当准确简要地概括公文的主要内容",
+            "doc_number": "发文字号由发文机关代字、年份和发文顺序号组成，年份用六角括号〔〕括入",
+            "date": "成文日期署会议通过或者发文机关负责人签发的日期，用阿拉伯数字将年、月、日标全",
+            "body": "正文是公文的主体，用来表述公文的内容，应当条理清楚、文字简练、用词准确"
+        }
+        
+        return hints.get(field, "")
+    
+    def generate_sample_content(self, doc_type: str) -> str:
+        """生成示例内容"""
+        samples = {
+            "通知": """
+为加强安全管理，防范安全事故发生，现就开展2026年度安全检查工作通知如下：
+
+一、检查时间
+2026年3月15日至2026年4月15日。
+
+二、检查范围
+公司各部门、各下属单位。
+
+三、检查内容
+（一）安全生产责任制落实情况；
+（二）安全管理制度执行情况；
+（三）安全隐患排查整改情况；
+（四）安全教育培训开展情况。
+
+四、工作要求
+（一）高度重视，精心组织；
+（二）认真检查，不留死角；
+（三）发现问题，及时整改；
+（四）总结经验，完善制度。
+
+请各部门、各单位按照本通知要求，认真组织开展安全检查工作，确保安全生产。
+
+特此通知。
+""",
+            "报告": """
+现将2025年度工作情况报告如下：
+
+一、主要工作完成情况
+（一）业务发展稳步推进；
+（二）重点项目顺利实施；
+（三）管理水平持续提升；
+（四）队伍建设不断加强。
+
+二、存在的问题和不足
+（一）业务创新能力有待提高；
+（二）人才队伍建设需要加强；
+（三）内部管理机制有待完善。
+
+三、2026年度工作计划
+（一）加快业务发展；
+（二）推进重点项目；
+（三）加强队伍建设；
+（四）完善管理机制。
+
+以上报告，请审阅。
+""",
+            "请示": """
+关于购置办公设备的请示
+
+因公司业务发展需要，现有办公设备已不能满足工作需求，拟购置一批办公设备。现请示如下：
+
+一、购置原因
+公司现有办公设备使用年限较长，性能下降，影响工作效率。
+
+二、购置清单
+（一）台式电脑10台；
+（二）笔记本电脑5台；
+（三）打印机3台；
+（四）复印机1台。
+
+三、预算金额
+预计总费用约15万元。
+
+四、资金来源
+拟从公司年度预算中列支。
+
+妥否，请批示。
+"""
+        }
+        
+        return samples.get(doc_type, "暂无示例内容")
+
+
+def create_smart_prompt_system():
+    """创建智能提示系统实例"""
+    return SmartPromptSystem()

package/skills/core/pdd-ba/SKILL.md

@@ -0,0 +1,305 @@
+---
+name: pdd-ba
+description: "PDD框架下的业务分析Skill，运用专业方法论进行需求推演和建模。当用户需要业务流程分析、需求建模或PRD审核时自动触发。支持中文触发：业务分析、需求分析、需求建模、5W1H分析。"
+license: MIT
+compatibility: 需要PRD文档或业务需求描述
+metadata:
+  author: "neuqik@hotmail.com"
+  version: "2.0"
+  parent: pdd-main
+  triggers:
+    - "/analyze" | "/audit" | "/doc"
+    - "分析业务流程" | "需求建模" | "PRD审核"
+    - "业务分析" | "5W1H" | "MECE"
+---
+
+# PDD-Business Analysis / 业务分析技能
+
+## Core Concept / 核心概念
+
+### 🇨🇳
+运用专业方法论进行需求分析和业务建模,输出结构化的业务分析结果,为后续功能点提取和开发规格生成提供基础。
+
+**输入**: PRD文档/业务需求描述/流程说明 | **输出**: 业务分析报告/用例图/流程图/状态图/CRUD矩阵 | **不负责**: 代码实现/测试编写/架构设计
+
+### 🇺🇸
+Applies professional methodology for requirement analysis and business modeling, outputting structured business analysis results as the foundation for subsequent feature extraction and development spec generation.
+
+**Input**: PRD documents / Business requirement descriptions / Process specifications | **Output**: Business analysis report / Use case diagrams / Flowcharts / State diagrams / CRUD matrix | **NOT responsible for**: Code implementation / Test writing / Architecture design
+
+## 方法论工具箱 / Methodology Toolbox
+
+### 🇨🇳
+### 5W1H 分析法
+
+| 维度 | 问题 | 分析要点 |
+|------|------|---------|
+| **Why** | 为什么做？ | 业务背景、目标、价值 |
+| **What** | 做什么？ | 功能范围、业务内容 |
+| **Who** | 谁来做？ | 角色划分、职责边界 |
+| **When** | 何时做？ | 时间要求、里程碑 |
+| **Where** | 在哪做？ | 使用场景、环境 |
+| **How** | 怎么做？ | 实现方式、流程 |
+
+### MECE 原则
+
+确保业务分析无遗漏、无重叠: 相互独立(各分析项不交叉重叠) | 完全穷尽(覆盖所有业务场景)
+
+**检验清单**: 子类别是否相互独立? | 所有类别是否覆盖完整? | 是否有遗漏的业务场景? | 是否有重叠的功能定义?
+
+### CRUD 矩阵
+
+分析实体与操作的对应关系(实体 × Create/Read/Update/Delete + 列表/导出)
+
+### 🇺🇸
+### 5W1H Analysis Method
+
+| Dimension | Question | Analysis Points |
+|-----------|----------|----------------|
+| **Why** | Why do it? | Business background, goals, value |
+| **What** | What to do? | Feature scope, business content |
+| **Who** | Who does it? | Role division, responsibility boundaries |
+| **When** | When to do? | Time requirements, milestones |
+| **Where** | Where to do? | Usage scenarios, environment |
+| **How** | How to do? | Implementation approach, process |
+
+### MECE Principle
+
+Ensure business analysis is complete and non-overlapping: Mutually Exclusive (analysis items don't overlap) | Collectively Exhaustive (covers all business scenarios)
+
+**Checklist**: Are sub-categories mutually exclusive? | Are all categories fully covered? | Any missing business scenarios? | Any overlapping feature definitions?
+
+### CRUD Matrix
+
+Analyze entity-operation mapping (entity × Create/Read/Update/Delete + List/Export)
+
+## 分析流程 / Analysis Flow
+
+### 🇨🇳
+### Step 1: 需求收集
+PRD文档原文 | 业务流程描述 | 业务规则文档 | 现有系统截图/文档 | 竞品分析资料 → 输出:原始需求清单
+
+### Step 2: 5W1H 分析
+
+对每个需求进行六维分析:
+- **Why**: 业务背景/核心价值/成功标准
+- **What**: 功能描述/范围边界/不做事项
+- **Who**: 发起角色/参与角色/审批角色/受益角色
+- **When**: 开始时间/结束时间/周期特性(实时/周期/事件)
+- **Where**: 使用场景/使用环境/接入方式
+- **How**: 实现方式/关键流程/技术约束
+
+### Step 3: 业务建模
+
+**a. 用例图(Use Case Diagram)**: 参与者+用例+关系
+
+**用例模板**:
+```markdown
+### UC-[编号]: [用例名称]
+**参与者**: [主要参与者]
+**前置条件**: [进入系统前的状态]
+**基本流程**: 1.[步骤1] → 2.[步骤2] → 3.[步骤3]
+**扩展流程**: [条件A]:[扩展1] | [条件B]:[扩展2]
+**异常流程**: [异常1]:[处理方式]
+```
+
+**b. 流程图(Process Flow)**: 开始→判断条件→分支处理→循环→结束
+
+**c. 状态图(State Diagram)**:
+
+**状态定义模板**:
+```markdown
+### 实体状态机
+**实体**: [实体名称]
+| 状态 | 代码 | 含义 | 可转入 |
+|------|------|------|--------|
+| 草稿 | DRAFT | 初始状态 | SUBMITTED |
+| 已提交 | SUBMITTED | 已提交待审核 | APPROVED,REJECTED |
+| 已审批 | APPROVED | 审批通过 | ARCHIVED |
+| 已拒绝 | REJECTED | 审批拒绝 | DRAFT |
+```
+
+### Step 4: CRUD 矩阵生成
+实体 × 创建/读取/更新/删除/列表/导出 的完整操作矩阵
+
+### Step 5: 业务规则提取
+
+| 规则ID | 规则描述 | 约束类型 | 优先级 |
+|--------|---------|---------|--------|
+| BR-001 | [规则描述] | 硬性/软性规则 | P0/P1/P2 |
+
+### 🇺🇸
+### Step 1: Requirement Gathering
+PRD document text | Business process descriptions | Business rule documents | Existing system screenshots/docs | Competitive analysis materials → Output: Raw requirement list
+
+### Step 2: 5W1H Analysis
+
+Six-dimensional analysis for each requirement:
+- **Why**: Business background / Core value / Success criteria
+- **What**: Feature description / Scope boundaries / Out of scope items
+- **Who**: Initiating role / Participating roles / Approval roles / Beneficiary roles
+- **When**: Start time / End time / Periodicity (real-time/cyclical/event)
+- **Where**: Usage scenarios / Usage environment / Access method
+- **How**: Implementation approach / Key processes / Technical constraints
+
+### Step 3: Business Modeling
+
+**a. Use Case Diagram**: Actors + Use cases + Relationships
+
+**Use Case Template**:
+```markdown
+### UC-[ID]: [Use Case Name]
+**Actors**: [Primary actor]
+**Preconditions**: [State before entering system]
+**Basic Flow**: 1.[Step 1] → 2.[Step 2] → 3.[Step 3]
+**Alternative Flows**: [Condition A]:[Alt 1] | [Condition B]:[Alt 2]
+**Exception Flows**: [Exception 1]:[Handling]
+```
+
+**b. Process Flow**: Start → Decision → Branch processing → Loop → End
+
+**c. State Diagram**:
+
+**State Definition Template**:
+```markdown
+### Entity State Machine
+**Entity**: [Entity Name]
+| State | Code | Meaning | Transitions To |
+|-------|------|---------|---------------|
+| Draft | DRAFT | Initial state | SUBMITTED |
+| Submitted | SUBMITTED | Pending review | APPROVED, REJECTED |
+| Approved | APPROVED | Review passed | ARCHIVED |
+| Rejected | REJECTED | Review rejected | DRAFT |
+```
+
+### Step 4: CRUD Matrix Generation
+Complete operation matrix: Entity × Create/Read/Update/Delete/List/Export
+
+### Step 5: Business Rule Extraction
+
+| Rule ID | Rule Description | Constraint Type | Priority |
+|---------|-----------------|-----------------|----------|
+| BR-001 | [Rule description] | Hard/Soft rule | P0/P1/P2 |
+
+## 输出规范 / Output Specification
+
+### 🇨🇳
+### 业务分析报告结构
+1. **概述**: 文档信息(模块编号/名称/日期/版本) | 业务背景 | 目标与价值
+2. **5W1H分析**: 详见Step 2
+3. **用例分析**: 详见Step 3.a
+4. **流程分析**: 详见Step 3.b
+5. **状态分析**: 详见Step 3.c
+6. **CRUD矩阵**: 详见Step 4
+7. **业务规则**: 详见Step 5
+8. **风险与假设**: 风险(ID/描述/影响/概率/应对) | 假设(ID/描述/验证方式)
+9. **附录**: 术语表 | 参考文档
+
+### 🇺🇸
+### Business Analysis Report Structure
+1. **Overview**: Document info (module number/name/date/version) | Business background | Goals & value
+2. **5W1H Analysis**: See Step 2
+3. **Use Case Analysis**: See Step 3.a
+4. **Process Analysis**: See Step 3.b
+5. **State Analysis**: See Step 3.c
+6. **CRUD Matrix**: See Step 4
+7. **Business Rules**: See Step 5
+8. **Risks & Assumptions**: Risks (ID/description/impact/probability/mitigation) | Assumptions (ID/description/verification method)
+9. **Appendix**: Glossary | Reference documents
+
+## Guardrails / 安全护栏
+
+### 🇨🇳
+**必须遵守**: 每个需求必须有明确的5W1H分析 | 每个业务实体必须有状态定义 | CRUD矩阵必须覆盖所有业务操作 | 业务规则必须标注优先级
+
+**避免事项**: ❌ 遗漏关键业务流程 | ❌ 状态定义不完整或不互斥 | ❌ 业务规则与实际业务不符 | ❌ 用例与流程描述不一致
+
+### 🇺🇸
+**MUST comply**: Every requirement MUST have complete 5W1H analysis | Every business entity MUST have state definitions | CRUD matrix MUST cover all business operations | Business rules MUST be marked with priority
+
+**Avoid**: ❌ Missing critical business flows | ❌ Incomplete or non-exclusive state definitions | ❌ Business rules not matching actual business | ❌ Inconsistency between use cases and process descriptions
+
+## 与其他技能协作 / Skill Collaboration
+
+### 🇨🇳
+| 协作技能 | 协作方式 | 传入数据 | 期望输出 |
+|---------|---------|---------|---------|
+| **pdd-main** | 流程调度 | 分析请求 | 业务分析报告 |
+| **pdd-extract-features** | Sequential | 分析报告 | 功能点矩阵 |
+| **pdd-generate-spec** | Sequential | 分析报告 | 开发规格 |
+
+### 🇺🇸
+| Collaborating Skill | Collaboration Mode | Input Data | Expected Output |
+|--------------------|-------------------|------------|-----------------|
+| **pdd-main** | Workflow scheduling | Analysis request | Business analysis report |
+| **pdd-extract-features** | Sequential | Analysis report | Feature matrix |
+| **pdd-generate-spec** | Sequential | Analysis report | Development spec |
+
+---
+
+## Iron Law / 核心铁律
+
+### 🇨🇳
+1. **方法论驱动分析**: 必须使用5W1H、MECE、CRUD等标准方法论进行分析,不得凭直觉或经验随意输出分析结果。
+2. **不越界做技术决策**: 业务分析的产出是业务层面的描述(用例、流程、规则),不涉及具体技术实现方案、数据库设计或接口定义。
+3. **完整性优先于速度**: 每个需求必须完成全部六个维度的5W1H分析,不得因为"看起来简单"而跳过某些维度。
+4. **可追溯性保障**: 分析报告中的每个结论都必须能追溯到PRD文档原文或用户提供的业务资料,不得编造不存在业务规则。
+5. **模型一致性**: 用例图、流程图、状态图之间必须保持一致,同一实体在不同图中的行为描述不得矛盾。
+
+**违规示例**: ❌ 在业务分析中直接设计数据库表结构 | ❌ 因为"登录功能很简单"而跳过其5W1H分析 | ❌ 编造PRD中未提及的业务规则 | ❌ 用例图中显示的操作在流程图中没有对应步骤 | ❌ 使用非标准术语而不在术语表中定义
+
+**合规示例**: ✅ 对每个需求执行完整的5W1H分析并记录到报告中 | ✅ 输出用例图和流程图时保持参与者与操作的一致性 | ✅ 所有业务规则都标注来源(如"源自PRD第3章第2节") | ✅ 在术语表中明确定义所有专业词汇 | ✅ 使用MECE原则检验功能点分类是否完整且互斥
+
+### 🇺🇸
+1. **Methodology-Driven Analysis**: MUST use standard methodologies (5W1H, MECE, CRUD) for analysis. NEVER output results based on intuition or experience alone.
+2. **No Technical Decision Overreach**: Business analysis outputs are business-level descriptions (use cases, processes, rules). Must NOT involve specific technical implementations, database designs, or API definitions.
+3. **Completeness Over Speed**: Every requirement MUST complete all six dimensions of 5W1H analysis. NEVER skip dimensions because something "looks simple".
+4. **Traceability Guarantee**: Every conclusion in the analysis report MUST be traceable to PRD document text or user-provided business materials. NEVER fabricate non-existent business rules.
+5. **Model Consistency**: Use case diagrams, flowcharts, and state diagrams MUST remain consistent. The same entity's behavior must not contradict across different diagrams.
+
+**Violation Examples**: ❌ Designing database table structures directly in business analysis | ❌ Skipping 5W1H analysis because "login is simple" | ❌ Fabricating business rules not mentioned in PRD | ❌ Operations in use case diagram have no corresponding steps in flowchart | ❌ Using non-standard terms without defining them in glossary
+
+**Compliance Examples**: ✅ Execute complete 5W1H analysis for each requirement and record in report | ✅ Maintain consistency between actors and operations when outputting use case and process diagrams | ✅ Cite sources for all business rules (e.g., "from PRD Ch.3 Sec.2") | ✅ Define all technical terms clearly in glossary | ✅ Use MECE principle to verify feature classification is complete and mutually exclusive
+
+---
+
+## Rationalization Table / 合理化防御表
+
+| # | Trap / 陷阱 | Question / 请问自己 | Action / 应该怎么做 |
+|---|-------------|------------------|------------------|
+| 1 | "这个需求很明显,不用写那么详细"<br>"This requirement is obvious, no need for detail" | 明显的需求也可能有隐含的业务规则和边界条件 | 至少完成最小化的5W1H分析模板,确保覆盖关键维度 |
+| 2 | "PRD已经写得很清楚了,我总结一下就行"<br>"PRD is clear enough, I'll just summarize" | 总结会丢失细节,后续技能需要完整的分析结果作为输入 | 按照标准模板逐项填写,而非概括性总结 |
+| 3 | "这个流程图画起来太费时间"<br>"Flowcharts take too much time" | 流程图是后续功能点提取的关键输入,缺失会导致遗漏 | 使用mermaid语法快速生成,确保至少覆盖主流程和异常分支 |
+| 4 | "状态定义差不多就行"<br>"State definitions are good enough" | 不完整的状态定义会导致状态机实现缺陷 | 确保每个状态都有明确的转入/转出条件和触发事件 |
+| 5 | "用户没提这个规则,但我认为应该有"<br>"User didn't mention this rule, but I think it should exist" | 编造业务规则会导致实现与实际需求不符 | 将推断标记为"假设"并请用户确认,不得直接作为确定规则 |
+
+**常见陷阱 / Common Traps**:
+1. **"表面分析"陷阱**: 只描述"做什么"而不深究"为什么"和"怎么做" → 强制完成5W1H全部分析,特别是Why和How维度
+2. **"技术前置"陷阱**: 过早引入技术实现细节 → 始终保持业务视角,技术决策留给后续技能
+3. **"模型脱节"陷阱**: 用例、流程、状态图各自独立绘制未进行交叉验证 → 在输出前强制执行模型一致性检查清单
+4. **"术语泛滥"陷阱**: 大量使用专业术语但不提供定义 → 强制维护术语表,首次出现时必须定义
+
+---
+
+## Red Flags / 红旗警告
+
+### Layer 1: Input Guards / 输入防护
+- **INPUT-BA-001**: PRD文档为空或内容少于100字 → 🔴 CRITICAL → 终止并提示用户提供有效的PRD文档 / Terminate and prompt user to provide valid PRD document
+- **INPUT-BA-002**: PRD文档格式无法解析(非文本/Markdown) → 🔴 CRITICAL → 提示支持的文档格式 / Prompt for supported document formats
+- **INPUT-BA-003**: 用户只提供了需求标题而无详细描述 → 🟡 WARN → 尝试通过对话收集更多信息,基于有限信息进行分析但标注不确定性 / Try to gather more info via conversation, analyze with limited data but mark uncertainty
+
+### Layer 2: Execution Guards / 执行防护
+- **EXEC-BA-001**: 跳过5W1H任一维度的分析 → 🔴 CRITICAL → 补充完整后再继续 / Complete missing dimensions before continuing
+- **EXEC-BA-002**: 业务规则未标注优先级(P0/P1/P2) → 🟡 WARN → 补充优先级标注,默认设为P1 / Add priority labels, default to P1
+- **EXEC-BA-003**: CRUD矩阵未覆盖所有已识别的实体 → 🔴 CRITICAL → 补充遗漏实体的CRUD分析 / Add CRUD analysis for missing entities
+- **EXEC-BA-004**: 状态机的状态转换缺少触发事件或条件 → 🟡 WARN → 补充转换条件说明 / Add transition condition descriptions
+
+### Layer 3: Output Guards / 输出防护
+- **OUTPUT-BA-001**: 业务分析报告缺少风险与假设章节 → 🔴 CRITICAL → 补充风险识别和假设声明 / Add risk identification and assumption statements
+- **OUTPUT-BA-002**: 报告中的结论无法追溯到源文档 → 🟡 WARN → 补充引用来源或标记为推断 / Add citation sources or mark as inference
+- **OUTPUT-BA-003**: 术语表中有未定义的术语 → 🟡 WARN → 补充术语定义或替换为通俗表达 / Define terms or replace with plain language
+
+**Trigger Handling / 触发处理流程:**
+🔴 CRITICAL → 立即停止,报告问题详情,等待指示 | Stop immediately & report details, wait for instructions
+🟡 WARN → 记录警告到分析日志,尝试自动修复,在最终报告中标注 | Log warning, attempt auto-fix, annotate in final report
+🔵 INFO → 记录信息,正常继续 | Log info, continue normally

package/skills/core/pdd-ba/_meta.json

@@ -0,0 +1 @@
+{"name": "pdd-ba", "version": "1.0.0", "category": "core", "description": "PDD framework business analysis Skill using professional methodology for requirement elicitation and modeling. Invoke when users input /analyze, /audit, /doc commands, or need business process analysis, requirement modeling, PRD review, stakeholder analysis, or 5W1H analysis. 支持中文触发：业务分析、需求分析、需求建模、5W1H分析。", "triggers": ["业务分析", "需求分析", "/analyze", "/audit", "需求建模", "5W1H分析", "business analysis", "requirement modeling", "5W1H analysis", "PRD review"]}

package/skills/core/pdd-ba/evals/default-evals.json

@@ -0,0 +1 @@
+[{"name": "SKILL.md-exists", "description": "技能文件存在且格式正确", "type": "structure", "expect": {"exists": true, "hasDescription": true}, "checks": {"file_exists": "SKILL.md"}}, {"name": "_meta.json-valid", "description": "元数据文件格式正确", "type": "structure", "expect": {"exists": true, "hasName": true, "hasVersion": true}, "checks": {"file_exists": "_meta.json"}}, {"name": "has-behavior-shaping", "description": "包含完整的行为塑造章节", "type": "quality", "expect": {"complete": true}, "contains": ["Iron Law", "Rationalization", "Red Flags"]}, {"name": "analysis-methods-defined", "description": "定义了业务分析方法论", "type": "content", "contains": ["分析", "方法论", "需求", "stakeholder"]}, {"name": "output-format-specified", "description": "明确了输出文档格式", "type": "content", "contains": ["输出", "格式", "模板", "template"]}]