claude-plan
Anthropic 風格的專業計畫管理 Skill - 自然語言驅動、預設並行、全自動 GitHub 整合
Best use case
claude-plan is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Anthropic 風格的專業計畫管理 Skill - 自然語言驅動、預設並行、全自動 GitHub 整合
Teams using claude-plan should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/claude-plan-majiayu000/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How claude-plan Compares
| Feature / Agent | claude-plan | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Anthropic 風格的專業計畫管理 Skill - 自然語言驅動、預設並行、全自動 GitHub 整合
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Claude Plan v1.0.0
> 自然語言描述 → **自動分解** → **預設並行** → PDCA 執行 → **全自動 GitHub** → Dashboard 追蹤 → 支持中斷續做
## 快速導覽
| 模組 | 用途 | 路徑 |
|------|------|------|
| **00-quickstart** | 快速開始 | [→](./00-quickstart/) |
| **01-planning** | 計畫制定(自然語言解析 + 任務分解) | [→](./01-planning/) |
| **02-tracking** | 進度追蹤(Git + Dashboard 混合) | [→](./02-tracking/) |
| **03-execution** | PDCA 執行循環 | [→](./03-execution/) |
| **04-interruption** | 中斷/恢復/智能插入 | [→](./04-interruption/) |
| **05-github** | GitHub 全自動整合 | [→](./05-github/) |
| **06-parallel** | 多 Session 預設並行 | [→](./06-parallel/) |
## 使用方式
```bash
/claude-plan [計畫描述]
# 範例
/claude-plan 建立電商網站 MVP
/claude-plan 重構認證系統,支援 OAuth 2.0
/claude-plan 撰寫 API 文檔並建立測試覆蓋率 80%
```
### Flags
```bash
--dashboard # 開啟視覺化 Dashboard
--add "需求" # 智能插入新需求
--resume # 從中斷處繼續(自動偵測)
--list # 列出所有進行中的計畫
--switch "專案" # 切換到指定專案
```
## 核心設計
### 基於 Anthropic 最佳實踐
| 來源 | 實踐 | 本 Skill 整合 |
|------|------|---------------|
| Boris Cherny Tips | 平行化策略(5-10 Claude 實例) | 預設並行,多 Session 自動認領 |
| Boris Cherny Tips | 驗證驅動(品質提升 2-3 倍) | PDCA Check 強制驗證 |
| Boris Cherny Tips | Subagents 策略 | 整合 verify-app, build-validator |
| Claude Code Docs | 自然語言命令 | 口語描述 → 自動分解任務 |
| Claude Code Docs | MCP 整合 | 連接 spec-workflow Dashboard |
### 設計原則
| 原則 | 說明 |
|------|------|
| **預設並行** | 任務自動分解為可並行單元,新 Session 自動 pick up |
| **全自動 GitHub** | commit → push → PR 無需干預 |
| **智能優先級** | AI 評估新需求緊急程度,自動排序 |
| **中斷友好** | 任何時刻中斷,下次自動從未完成處繼續 |
## 執行流程
```
/claude-plan 建立電商 MVP
↓
┌─────────────────────────────────────────────────────────┐
│ Phase 1: 自然語言解析 │
│ • 理解需求意圖 │
│ • 識別子目標 │
│ • 定義完成標準 │
└─────────────────────────────────────────────────────────┘
↓
┌─────────────────────────────────────────────────────────┐
│ Phase 2: 任務自動分解 │
│ • 分解為可並行單元 │
│ • 標記依賴關係 │
│ • 估算複雜度 │
└─────────────────────────────────────────────────────────┘
↓
┌─────────────────────────────────────────────────────────┐
│ Phase 3: 並行就緒 │
│ ┌─────────┬─────────┬─────────┐ │
│ │ Task A │ Task B │ Task C │ ← 可並行 │
│ │ [待認領] │ [待認領] │ [待認領] │ │
│ └─────────┴─────────┴─────────┘ │
│ ↓ │
│ ┌─────────┐ │
│ │ Task D │ ← 依賴 A+B │
│ │ [阻塞中] │ │
│ └─────────┘ │
└─────────────────────────────────────────────────────────┘
↓
┌─────────────────────────────────────────────────────────┐
│ Phase 4: PDCA 執行(每個 Task) │
│ Plan → Do → Check → Act │
│ 驗證通過 → 自動 commit + push │
└─────────────────────────────────────────────────────────┘
↓
┌─────────────────────────────────────────────────────────┐
│ Phase 5: 全自動 GitHub │
│ • 每個 Task 完成 → commit │
│ • Milestone 完成 → push + PR │
└─────────────────────────────────────────────────────────┘
```
## 狀態存儲
### 混合模式架構
```
.claude/plans/ # Git-tracked(主計畫)
├── index.json # 所有計畫索引
└── ecommerce-mvp/
├── plan.md # 主計畫文件
├── north-star.md # 北極星錨定
├── sessions/
│ ├── session-001.json # Session 狀態快照
│ └── session-002.json
└── tasks/
├── backend-api.md # 子任務詳情
└── frontend-ui.md
specs/ecommerce-mvp/ # spec-workflow 整合(Dashboard)
├── requirements.md
├── design.md
├── tasks.md
└── implementation-log/
```
## 多 Session 並行
### 運作機制
```
新 Session 啟動: /claude-plan
↓
┌─────────────────────────────────────────────────────────┐
│ 1. 讀取 .claude/plans/index.json │
│ 2. 檢查可認領的任務(無鎖 + 依賴已完成) │
│ 3. 自動認領第一個可用任務 │
│ 4. 創建任務鎖(防止其他 Session 重複認領) │
│ 5. 執行 PDCA │
│ 6. 完成後釋放鎖 + 更新狀態 │
└─────────────────────────────────────────────────────────┘
```
### 任務鎖機制
```json
{
"task_id": "backend-api",
"locked_by": "session-001",
"locked_at": "2025-01-21T10:30:00Z",
"expires_at": "2025-01-21T11:30:00Z"
}
```
## Dashboard 整合
```bash
/claude-plan --dashboard
```
**功能**:
- 視覺化進度(spec-workflow Dashboard)
- 即時 Session 狀態
- 任務依賴圖
- GitHub 活動追蹤
## 智能需求插入
```bash
/claude-plan --add "新增支付功能"
```
**AI 評估流程**:
1. 分析新需求內容
2. 評估緊急程度(P0-P3)
3. 檢查與現有任務的依賴
4. 自動決定:插隊 or 排隊
5. 更新計畫 + 通知相關 Session
## 中斷與恢復
### 自動狀態保存
每次中斷時自動保存:
- 當前任務進度
- 已完成的驗證
- 未提交的變更
### 恢復機制
```bash
/claude-plan --resume
# 或直接
/claude-plan # 自動偵測未完成計畫
```
## 完成信號
- `✅ PLAN COMPLETED: [計畫名稱]`
- `⏸️ NEED HUMAN: [原因]`
- `🔄 TASKS REMAINING: X/Y`
## 相關資源
- [Boris Cherny Claude Code Tips](../.claude/memory/learnings/2025-01-07-boris-cherny-claude-code-tips.md)
- [spec-workflow 整合](./05-integration/_base/spec-workflow.md)
- [Anthropic Claude Code 官方文檔](https://code.claude.com/docs)Related Skills
developing-claude-code-plugins
Build, test, and distribute Claude Code plugins with slash commands, agents, skills, hooks, MCP servers, and LSP servers. MUST be loaded when creating, reviewing, debugging, or distributing plugins. Use PROACTIVELY when user mentions plugins, extensions, marketplaces, or wants to add custom commands/agents to Claude Code.
claude-skill-creator
Guide for creating effective Claude Code skills with proper YAML frontmatter, directory structure, and best practices. Use when creating new skills, updating existing skills, or learning about skill development.
claude-settings-audit
Analyze a repository to generate recommended Claude Code settings.json permissions. Use when setting up a new project, auditing existing settings, or determining which read-only bash commands to allow. Detects tech stack, build tools, and monorepo structure.
claude-md-editor
Edit and maintain CLAUDE.md and project documentation files. Use when modifying documentation, asking about best practices, or performing documentation maintenance tasks.
claude-improve-config
Self-reflect on the current session to identify mistakes and propose improvements to .claude configuration (CLAUDE.md, hooks, skills).
claude-hooks-reference-2026
Complete reference for Claude Code hooks system (January 2026). Use when creating hooks, understanding hook events, matchers, exit codes, JSON output control, environment variables, plugin hooks, or implementing hook scripts.
claude-ecosystem
Claude Code ecosystem expertise. Modules: CLI tool (setup, slash commands, MCP servers, hooks, plugins, CI/CD), extensibility (agents, skills, output styles creation), CLAUDE.md (project instructions, optimization). Actions: configure, troubleshoot, create, deploy, integrate, optimize Claude Code. Keywords: Claude Code, Anthropic, CLI tool, slash command, MCP server, Agent Skill, hook, plugin, CI/CD, enterprise, CLAUDE.md, agentic coding, agent, skill, output-style, SKILL.md, subagent, Task tool, project instructions, token optimization. Use when: learning Claude Code features, configuring settings, creating skills/agents/hooks, setting up MCP servers, troubleshooting issues, CI/CD integration, initializing or optimizing CLAUDE.md files.
claude-code
Claude Code CLI and development environment. Use for Claude Code features, tools, workflows, MCP integration, configuration, and AI-assisted development.
claude-code-slash-commands
This skill should be used when the user asks to "create a command", "write a slash command", "build a plugin command", or wants to add custom commands to Claude Code.
claude-code-meta
Build Claude Code extensions - skills, agents, hooks, plugins, marketplaces, slash commands. Use when creating Claude Code components, building new skills, writing agents, creating hooks, making plugins, setting up marketplaces, writing slash commands, fixing extension configurations, or troubleshooting Claude Code extensions. Includes builder agents for autonomous creation. Not for looking up docs - use claude-code-docs-reference for that.
claude-code-analyzer
Analyzes Claude Code usage patterns and provides comprehensive recommendations. Runs usage analysis, discovers GitHub community resources, suggests CLAUDE.md improvements, and fetches latest docs on-demand. Use when user wants to optimize their Claude Code workflow, create configurations (agents/skills/commands), or set up project documentation.
agentsmd-claudemd-generator
Generate AGENTS.md and CLAUDE.md files for a repository. AGENTS.md provides cross-tool agent instructions (supported by Claude Code, Cursor, Windsurf, Zed, Codex, and others). CLAUDE.md adds Claude-specific configuration and references AGENTS.md via @import. Use when a repo needs agent onboarding or when starting a new project.