overall-designer

Original：🇺🇸 English

Translated

Generate high-level overall design for large features. Produces database model overview, core flow diagrams, and phased iteration plan. Use when user needs overall architecture planning for a big feature, not detailed class-level design. Triggers on "总体设计", "初步设计", "方案设计", "overall design". Does NOT do detailed design — use detail-designer for class/method level. Does NOT implement code.

1installs

Sourcejoesmart/solopreneur-skills

Added on2026-03-14

NPX Install

npx skill4agent add joesmart/solopreneur-skills overall-designer

SKILL.md Content

View Translation Comparison →

Overall Design Generator

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.

Step 1: Understand & Clarify

1.1 Decompose Requirement

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"

1.2 Explore Current System

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)

1.3 Clarify with User (mandatory)

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

1.4 Verify Understanding

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

Step 2: Design

2.1 Database Model Overview

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]

2.2 Core Flow Diagrams

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 ★.

2.3 Data Trace (internal only)

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

Step 3: Phased Iteration Plan (Core Output)

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.

Output Template

markdown

# [Feature] - 初步设计方案
## 一、背景与问题
## 二、总体方案 (overview diagram + design principles)
## 三、数据库模型概要
## 四、核心流程链路 (ASCII flows)
## 五、迭代实施计划 (Phases + dependency diagram)
## 六、关键设计决策 (table: decision + rationale)
## 七、待确认事项

Thinking Discipline

After each step, apply thinking discipline.