overall-designer

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Overall Design Generator

总体设计生成器

Output: Background + Database Overview + Core Flow + Phased Iteration Plan.

Before starting: Check if

PROJECT_MAP.md

exists at project root. If yes, read it first.

输出内容：背景 + 数据库概览 + 核心流程 + 分阶段迭代计划。

开始前：检查项目根目录是否存在

PROJECT_MAP.md

文件。如果存在，请先阅读该文件。

Step 1: Understand & Clarify

步骤1：理解与澄清

1.1 Decompose Requirement

1.1 需求拆解

Parse requirement, extract every business concept and action verb
For each concept: does it exist in current system? Which table/service/module?
For each action: which systems involved? Data flow?
Mark: "known" / "to investigate (code)" / "need user confirmation"

解析需求，提取所有业务概念和动作动词
针对每个概念：它是否存在于现有系统中？对应哪个表/服务/模块？
针对每个动作：涉及哪些系统？数据流是怎样的？
标记："已知" / "需调研（代码）" / "需用户确认"

1.2 Explore Current System

1.2 探索现有系统

bash

SHARED="$(dirname "${CLAUDE_SKILL_DIR}")/_shared/scripts"
bash "$SHARED/detect-project-type.sh" .
bash "$SHARED/detect-tech-stack.sh" .
bash "$SHARED/scan-dependencies.sh" .

Find related existing features, trace data flows
Identify reusable capabilities (existing APIs, services, queues)

bash

SHARED="$(dirname "${CLAUDE_SKILL_DIR}")/_shared/scripts"
bash "$SHARED/detect-project-type.sh" .
bash "$SHARED/detect-tech-stack.sh" .
bash "$SHARED/scan-dependencies.sh" .

查找相关的现有功能，追踪数据流
识别可复用的能力（现有API、服务、队列）

1.3 Clarify with User (mandatory)

1.3 与用户确认（必填）

Compile all "need confirmation" questions, prioritize
Ask the user, wait for answers before proceeding
If answers raise new questions → keep asking

整理所有“需确认”的问题，并按优先级排序
询问用户，等待回复后再继续
如果回复引发新问题 → 继续询问

1.4 Verify Understanding

1.4 验证理解

Can you explain the full picture: what problem, which systems, how data flows?
Can you list the core challenges/risks?
If not → keep exploring.

你能否完整阐述整体情况：要解决什么问题、涉及哪些系统、数据流如何？
你能否列出核心挑战/风险？
如果不能 → 继续探索。

Step 2: Design

步骤2：设计

2.1 Database Model Overview

2.1 数据库模型概览

For each new/modified table:

Table name + one-sentence purpose
Core business fields only (name + type + description)

Table relationships diagram:

b_lead (1) ──── (1) b_lead_element
   │
   └── sany_lead_id ──── inquiry_id (inquiry-service)

Mark: [New] / [Modify existing]

针对每个新增/修改的表：

表名 + 一句话用途说明
仅包含核心业务字段（名称 + 类型 + 描述）

表关系图：

b_lead (1) ──── (1) b_lead_element
   │
   └── sany_lead_id ──── inquiry_id (inquiry-service)

标记：[新增] / [修改现有]

2.2 Core Flow Diagrams

2.2 核心流程图

For each key path (normal flow, async callback, error flow):

┌──────────────────────────────────────────────────────┐
│ [System]: [Module] [Status: New/Modify/Existing]     │
├──────────────────────────────────────────────────────┤
│  Responsibility: one sentence                        │
│  Input: key data     Output: key data                │
│  Steps:                                              │
│      ├── Step 1: description                         │
│      └── Step 2: description                         │
└────────────────────────┬─────────────────────────────┘
                         │ [HTTP/MQ/Callback/Method call]
                         ▼

Mark new/changed parts with ★.

针对每个关键路径（正常流程、异步回调、错误流程）：

┌──────────────────────────────────────────────────────┐
│ [System]: [Module] [Status: New/Modify/Existing]     │
├──────────────────────────────────────────────────────┤
│  Responsibility: one sentence                        │
│  Input: key data     Output: key data                │
│  Steps:                                              │
│      ├── Step 1: description                         │
│      └── Step 2: description                         │
└────────────────────────┬─────────────────────────────┘
                         │ [HTTP/MQ/Callback/Method call]
                         ▼

用★标记新增/修改的部分。

2.3 Data Trace (internal only)

2.3 数据追踪（内部使用）

Trace 2+ data sets through flows. Fix design issues. Not in output.

通过流程追踪2组以上的数据集。修复设计问题。不包含在输出内容中。

Step 3: Phased Iteration Plan (Core Output)

步骤3：分阶段迭代计划（核心输出）

Split by end-to-end verifiable features, not by tech layers.

Each Phase:

Number and name
Goal: one sentence
Deliverable: what user/system can do after this Phase
Systems involved
Steps: system, change summary, verification method, dependencies
★ Milestone: one sentence

Dependency diagram:

Phase 1: [name]
  ├── Step 1.1 ← no deps
  ├── Step 1.2 ← Step 1.1
  └── Step 1.3 ← Step 1.2
  ★ Milestone: [deliverable]
       │
       ▼
Phase 2: [name] ...

Verify: all DB models and flows are covered by Phases.

按可端到端验证的功能拆分，而非按技术层拆分。

每个阶段包含：

阶段编号与名称
目标：一句话说明
交付成果：本阶段完成后，用户/系统可实现的功能
涉及系统
步骤：系统、变更摘要、验证方法、依赖项
★ 里程碑：一句话说明

依赖关系图：

Phase 1: [name]
  ├── Step 1.1 ← no deps
  ├── Step 1.2 ← Step 1.1
  └── Step 1.3 ← Step 1.2
  ★ Milestone: [deliverable]
       │
       ▼
Phase 2: [name] ...

验证：所有数据库模型和流程均被各阶段覆盖。

Output Template

输出模板

markdown

undefined

markdown

undefined

[Feature] - 初步设计方案

[功能名称] - 初步设计方案

一、背景与问题

二、总体方案 (overview diagram + design principles)

二、总体方案 (概览图 + 设计原则)

三、数据库模型概要

四、核心流程链路 (ASCII flows)

四、核心流程链路 (ASCII流程图)

五、迭代实施计划 (Phases + dependency diagram)

五、迭代实施计划 (阶段 + 依赖关系图)

六、关键设计决策 (table: decision + rationale)

六、关键设计决策 (表格：决策 + 理由)

七、待确认事项

undefined

undefined

Thinking Discipline

思考规范

After each step, apply thinking discipline.

每个步骤完成后，应用思考规范。