💡 摘要
Agentic Startup 是一个以规范驱动的开发框架，通过并行代理执行来提高编码效率。
🎯 适合人群
🤖 AI 吐槽: “看起来很能打，但别让配置把人劝退。”
风险：Medium。建议检查：是否执行 shell/命令行指令；是否发起外网请求（SSRF/数据外发）；API Key/Token 的获取、存储与泄露风险；文件读写范围与路径穿越风险。以最小权限运行，并在生产环境启用前审计代码与依赖。
Table of Contents
- 🤖 What is The Agentic Startup?
- 🚀 Quick Start
- 📖 The Complete Workflow
- 🎯 Which Command Should I Use?
- 📦 Plugins
- 🎨 Output Styles
- 📊 Statusline
- 🎯 Philosophy
- 📚 Documentation
🤖 What is The Agentic Startup?
The Agentic Startup is a spec-driven development framework for Claude Code. Create comprehensive specifications before coding, then execute with parallel specialist agents—Y Combinator energy meets engineering discipline.
Key Features:
- Native Claude Code Integration — Marketplace plugins with zero configuration
- Spec-Driven Development — PRD → SDD → Implementation Plan → Code
- Parallel Agent Execution — Multiple specialists work simultaneously
- Quality Gates — Built-in validation at every stage
Installation
Requirements: Claude Code v2.0+ with marketplace support
curl -fsSL https://raw.githubusercontent.com/rsmdt/the-startup/main/install.sh | sh
This installs the core plugins, configures the default output style, and sets up the statusline with a customizable config file.
# Add The Agentic Startup marketplace /plugin marketplace add rsmdt/the-startup # Install the Start plugin (core workflows) /plugin install start@the-startup # (Optional) Install the Team plugin (specialized agents) /plugin install team@the-startup
After installation:
# (Optional) Create project governance rules /start:constitution # Auto-enforced during specify, implement, review # Switch output styles anytime /output-style start:The Startup # High-energy, fast execution (default) /output-style start:The ScaleUp # Calm confidence, educational
🚀 Quick Start
Create a specification and implement it:
# Create a specification /start:specify Add user authentication with OAuth support # Execute the implementation /start:implement 001
That's it! You're now using spec-driven development.
📖 The Complete Workflow
The Agentic Startup follows spec-driven development: comprehensive specifications before code, ensuring clarity and reducing rework.
All Commands at a Glance
┌──────────────────────────────────────────────────────────┐
│ SETUP (optional) │
│ │
│ /start:constitution ► Create project governance rules │
│ (auto-enforced in BUILD workflow) │
└──────────────────────────────────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────┐
│ BUILD (primary flow) │
│ │
│ /start:specify ────► Create specs (PRD + SDD + PLAN) │
│ │ ↳ Constitution checked on SDD │
│ ▼ │
│ /start:validate ───► Check quality (3 Cs framework) │
│ │ ↳ Constitution mode available │
│ ▼ │
│ /start:implement ──► Execute plan phase-by-phase │
│ │ ↳ Constitution + drift enforced │
│ ▼ │
│ /start:review ─────► Multi-agent code review │
│ │ ↳ Constitution compliance checked │
│ ▼ │
│ /start:document ───► Generate/sync documentation │
└──────────────────────────────────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────┐
│ MAINTAIN (as needed) │
│ │
│ /start:analyze ────► Discover patterns & rules │
│ │
│ /start:refactor ───► Improve code (preserve behavior) │
│ │
│ /start:debug ──────► Fix bugs (root cause analysis) │
└──────────────────────────────────────────────────────────┘
Step-by-Step Walkthrough
Step 1: Create Your Specification
/start:specify Add real-time notification system with WebSocket support
This creates a specification directory with three documents:
docs/specs/001-notification-system/
├── product-requirements.md # What to build and why
├── solution-design.md # How to build it technically
└── implementation-plan.md # Executable tasks and phases
The spec cycle may take 15-30 minutes. Claude will research your codebase, ask clarifying questions, and produce comprehensive documents. The process naturally involves multiple back-and-forth exchanges.
Step 2: Handle Context Limits (Resume Pattern)
Large specifications may approach Claude's context window limits. When this happens:
# Start a new conversation and resume where you left off /start:specify 001
The resume pattern:
- Pass the spec ID (e.g.,
001) instead of a description - Claude reads the existing spec files and continues from there
- You can reset context as many times as needed
- Each document (PRD → SDD → PLAN) can be completed in separate sessions if needed
Pro tip: If Claude suggests "you may want to reset context", do it! The quality of output improves with fresh context.
Step 3: Validate Before Implementation
/start:validate 001
This quality gate checks:
- Completeness - All sections filled, no missing details
- Consistency - No contradictions between documents
- Correctness - Requirements are testable and achievable
Validation is advisory—it provides recommendations but doesn't block you.
Step 4: Execute the Implementation
/start:implement 001
Claude will:
- Parse the implementation plan
- Execute phases sequentially (with your approval between phases)
- Run tests after each task
- Use parallel agents within phases for speed
Large implementations may also need context resets. Simply run /start:implement 001 again in a fresh conversation—Claude tracks progress in the spec files.
Step 5: Review and Ship
/start:review
Four parallel specialists review your code:
- 🔒 Security - Authentication, authorization, input validation
- ⚡ Performance - Query optimization, memory management
- ✨ Quality - Code style, design patterns, maintainability
- 🧪 Tests - Coverage gaps, edge cases
🎯 Which Command Should I Use?
Decision Tree
What do you need to do?
│
├─ Want project-wide guardrails? ─────────► /start:constitution
│
├─ Build something new? ──────────────────► /start:specify
│ Then: /start:validate → /start:implement
│
├─ Understand existing code? ─────────────► /start:analyze
│ └─ Want to improve it? ───────────────► Then: /start:refactor
│
├─ Something is broken? ──────────────────► /start:debug
│
├─ Code ready for merge? ─────────────────► /start:review
│
├─ Need documentation? ───────────────────► /start:document
│
└─ Check constitution compliance? ────────► /start:validate constitution
Command Reference
| Command | Purpose | When to Use |
|---------|---------|-------------|
| /start:constitution | Create governance rules | Establish project-wide guardrails |
| /start:specify | Create specifications | New features, complex changes |
| /start:implement | Execute plans | After spec is validated |
| /start:validate | Check quality | Before implementation, after specs |
| /start:review | Multi-agent code review | Before merging PRs |
| /start:document | Generate documentation | After implementation |
| /start:analyze | Extract knowledge | Understanding existing code |
| /start:refactor | Improve code quality | Cleanup without behavior change |
| /start:debug | Fix bugs | When something is broken |
Capability Matrix
| Capability | constitution | specify | implement | validate | review | document | analyze | refactor | debug | |------------|:------------:|:-------:|:---------:|:--------:|:------:|:--------:|:-------:|:--------:|:-----:| | Creates specifications | - | ✅ | - | - | - | - | - | - | - | | Executes implementation plans | - | - | ✅ | - | - | - | - | - | - | | Runs tests | - | - | ✅ | ✅ | - | - | - | ✅ | ✅ | | Creates git branches | - | ✅ | ✅ | - | - | - | - | ✅ | - | | Creates PRs | - | ✅ | ✅ | - | - | - | - | - | - | | Multi-agent parallel | - | ✅ | ✅ | - | ✅ | ✅ | ✅ | - | - | | **Security
优点
- 简化的开发流程
- 并行执行提高速度
- 内置质量检查
- 与 Claude Code 集成
缺点
- 需要 Claude Code v2.0 及以上版本
- 新用户的学习曲线
- 依赖外部插件
- 管理规范的潜在复杂性
相关技能
免责声明：本内容来源于 GitHub 开源项目，仅供展示和评分分析使用。
版权归原作者所有 rsmdt.