CSnCompany_2-0 — An AI Team You Hire Inside Claude Code
🇺🇸 English · 🇰🇷 한국어
TL;DR — One marketplace, eleven AI teammates. Install it once, and you can call a CEO, PM, Architect, Designer, QA Engineer, Code Reviewer, and DevOps engineer from inside Claude Code with simple slash commands like /cs-ceo or /CS-test.
🤔 What is this? (For Beginners)
Claude Code is Anthropic's official AI coding CLI. It supports plugins — bundles of slash commands, agents, and skills you install on top of it.
CSnCompany_2-0 is a marketplace that bundles 11 plugins, each one a specialist on a virtual AI team:
You ──▶ /cs-ceo "build a dashboard"
│
▼
🧭 CEO decides which teammates to call
│
┌──────────┼──────────┬──────────┐
▼ ▼ ▼ ▼
🏗️ Plan 🎨 Design 🧪 Test 🚢 Ship
You don't need to remember which command does what — type /cs-ceo "your goal" and it dispatches the right teammates. Or call them directly when you know what you need.
👥 The Team
| Member | Plugin | Slash Command | What it does |
|---|
| 🧭 CEO | cs-ceo | /cs-ceo "goal" | Estimates effort, picks teammates, dispatches them. Start here if unsure. |
| 💬 PM | cs-clarify | /cs-clarify | Asks Socratic questions, surfaces hidden assumptions, prevents over-engineering |
| 🏗️ Architect | CS-plan | /CS-plan "feature" | TDD + Clean Architecture plan: domain analysis, architecture, test strategy, checklist |
| 🎨 Designer | cs-design | /cs-design <url> | 5-agent design review: hierarchy, interaction, design system, a11y, anti-patterns |
| 🎨 Design Reference | cs-design-sample1 | /cs-design-sample1 | Crextio-style design guide for Tailwind/Next.js dashboards |
| 🧪 QA Engineer | CS-test | /CS-test <url> | 14-agent web test: security, SEO, perf, a11y, DB, PWA, touch, image |
| 🔍 Code Reviewer | CS-codebase-review | /CS-codebase-review ./src | 5-agent review: architecture, quality, security, perf, maintainability |
| 🚢 DevOps | cs-ship | /cs-ship | Pre-PR validation: spec compliance, coverage, commit messages |
| ⚡ Team Lead | cs-smart-run | /cs-smart-run "task" | Plan with Opus → execute with Sonnet agents in parallel |
| 📚 Knowledge Keeper | cs-experiencing | /cs-experiencing | Versioned learnings + /cs-end session-wrap (author-only push) |
| 🗣️ Language Coach | convo-maker | /convo-maker | Turns session Q&A into natural American English conversations |
🚀 Install in 60 seconds
Prerequisite
Install Claude Code:
npm install -g @anthropic-ai/claude-code
Then launch it:
claude
Step 1 — Add the marketplace
Inside Claude Code, paste:
/plugin marketplace add intenet1001-commits/CSnCompany_2-0
Step 2 — Install the plugins you want
Pick à la carte, or install everything:
/plugin install cs-ceo@CSnCompany_2-0
/plugin install cs-clarify@CSnCompany_2-0
/plugin install CS-plan@CSnCompany_2-0
/plugin install cs-design@CSnCompany_2-0
/plugin install cs-design-sample1@CSnCompany_2-0
/plugin install CS-test@CSnCompany_2-0
/plugin install CS-codebase-review@CSnCompany_2-0
/plugin install cs-ship@CSnCompany_2-0
/plugin install cs-smart-run@CSnCompany_2-0
/plugin install cs-experiencing@CSnCompany_2-0
/plugin install convo-maker@CSnCompany_2-0
Step 3 — Restart Claude Code
That's it. Type / in Claude Code and you'll see the new commands.
🧭 Don't know where to start?
Ask the CEO:
/cs-ceo "I want to build a user dashboard with auth"
The CEO estimates effort, decides which teammates to call (PM, Architect, Designer, etc.), and runs them in the right order. You just sit back and answer when they ask clarifying questions.
💡 Common Workflows
Build a new feature from scratch
/cs-clarify "add Stripe payments" # PM: surface assumptions
↓
/CS-plan "Stripe checkout + webhook" # Architect: TDD plan
↓
… you implement code …
↓
/CS-test https://staging.example.com # QA: 14-agent web test
↓
/CS-codebase-review ./src # Reviewer: 5-agent code review
↓
/cs-ship # DevOps: pre-PR gate
Audit an existing site
/cs-design https://example.com # Visual + UX review
/CS-test https://example.com # Security/SEO/perf/a11y
Just let the CEO drive
/cs-ceo "audit my landing page and tell me what to fix first"
🏛️ Architecture — Lead-Agent Pattern
Every multi-agent plugin uses the lead-agent pattern: the main conversation spawns one lead agent, and the lead orchestrates N specialist workers internally. Worker output never pollutes your main context — only the final synthesized report comes back.