시작하기 전에 반드시 이해해야 할 한 가지

BALIA OS에서 당신은 CEO입니다.

OpenClaw에는 main이라는 기본 에이전트가 있습니다. 이 에이전트는 건드리거나 설정하거나 사용하지 않습니다. 백그라운드에서 그냥 두면 됩니다. 명령을 내리는 것은 바로 당신입니다.

구조는 단순합니다:

INFO

You (CEO) ↓ direct commands via VS Code or OpenGoat UI ├── Agent #1 (Operations) → reports to #agent1-channel in Discord ├── Agent #2 (Marketing) → reports to #agent2-channel in Discord ├── Agent #3 (Finance) → reports to #agent3-channel in Discord ├── Agent #4 (Specialist A) → reports to #agent4-channel in Discord ├── Agent #5 (Specialist B) → reports to #agent5-channel in Discord └── Agent #6 (Security) → reports to #security-channel in Discord

중간 관리자도 없고, 당신과 팀 사이에 오케스트레이터 에이전트도 없습니다. 당신이 각 에이전트에게 직접 말합니다. 각 에이전트는 자신만의 Discord 채널에 직접 보고합니다. 당신은 보고서를 읽습니다.

이것이 실제로 작동하는 설계입니다. 단순하고 직접적이며, 서브에이전트 오류가 없습니다.

---

이 섹션의 진행 방식

아래 프롬프트를 복사하여 Claude, ChatGPT, 또는 Gemini에 붙여넣으세요. AI가 한 번에 하나씩 질문하며 단계별로 팀 설계를 안내하고, 결정 사항을 함께 검토하면서 팀 프로필을 완성해 나갑니다.

대화가 끝나면 AI가 완성된 팀 설계 문서를 출력합니다. 이 문서를 저장해 두세요. 섹션 3에서 에이전트 파일을 생성할 때 사용합니다.

잘 모르는 부분이 있다면 그냥 말하세요. AI가 함께 해결해 드립니다.

---

▶ 이것을 AI에 붙여넣어 시작하세요

You are helping me set up BALIA OS — a multi-agent AI operating system 
where I configure a team of AI agents to run my business.

IMPORTANT STRUCTURE TO UNDERSTAND:
- I am the CEO. I give commands directly to each agent.
- There is no middle manager agent. No orchestrator between me and my team.
- Each agent reports to their own dedicated Discord channel.
- I read the reports and give the next command.
- OpenClaw has a default "main" agent — we leave it alone and don't use it.

My team has 6 slots:
- Slots #1 through #5: I configure these for my business
- Slot #6: Always Security — monitors threats, quarantines new cassettes, 
  protects API keys. This slot never changes.

Your job is to act as my setup concierge. Guide me through designing 
my team step by step. Ask me one question at a time. Wait for my answer 
before moving on. If my answer is vague, ask a follow-up to make it 
more specific.

Here is the process we will follow:

STEP 1 — My Business Tasks
Ask me: "What does your business actually need people to do every day? 
Just describe it in your own words — don't think about AI yet."
Help me turn my answer into a clear task list.

STEP 2 — Grouping Into Roles
Group my tasks into 5 roles (Slot #1 through #5).
Show me the groupings and ask if they feel right. 
Adjust based on my feedback.
For each role, also suggest a Discord channel name 
(e.g. #operations, #marketing, #finance).

STEP 3 — Naming Each Agent
For each role, suggest 3 name options. 
Explain the feel of each name in one sentence.
Ask me which name I want, or if I have my own idea.
Names should be short, distinct from each other, and easy to remember.

STEP 4 — Mission Statements
For each agent, write a 1-2 sentence mission statement that defines:
- What they are responsible for
- What they are NOT responsible for
Ask me if each one feels right. Adjust based on my feedback.

STEP 5 — Personality
For each agent, suggest a personality profile based on their role.
Ask me to confirm or adjust:
- Tone (formal / casual / direct / warm)
- Communication style (concise / detailed / analytical)
- Language (English / Japanese / Chinese / multilingual)

FINAL OUTPUT
Once all 5 steps are complete, produce a completed Team Design Document 
in this exact format:

---
TEAM DESIGN DOCUMENT
Business: [my business name or description]
Date: [today's date]

HOW I GIVE COMMANDS:
- I (CEO) give commands directly to each agent via VS Code or OpenGoat UI
- Each agent reports to their own Discord channel
- No middle manager agent

SLOT #1 — OPERATIONS
Name: 
Role title: 
Mission: 
Personality: 
Language: 
Discord channel: 

SLOT #2 — MARKETING & OUTREACH
Name: 
Role title: 
Mission: 
Personality: 
Language: 
Discord channel: 

SLOT #3 — FINANCE & REPORTING
Name: 
Role title: 
Mission: 
Personality: 
Language: 
Discord channel: 

SLOT #4 — SPECIALIST A
Name: 
Role title: 
Mission: 
Personality: 
Language: 
Discord channel: 

SLOT #5 — SPECIALIST B
Name: 
Role title: 
Mission: 
Personality: 
Language: 
Discord channel: 

SLOT #6 — SECURITY
Name: Security Guardian (fixed — do not change)
Role title: Security Monitor
Mission: Threat scanning every 4 hours. Quarantine all new cassettes 
before activation. Monitor API keys and webhooks continuously.
Personality: 
Language: 
Discord channel: #security
---

Start now. Ask me the first question.

---

2단계 — AI 컨시어지와 대화하기

AI 컨시어지와 함께 질문을 하나씩 풀어나가세요. 몇 가지 팁을 드립니다:

• 어떤 역할이 필요한지 모르겠다면, 달성하고 싶은 목표를 설명하세요 — AI가 구조를 제안해 드립니다
• 제안된 이름이 마음에 들지 않는다면, 그렇게 말하세요 — AI가 다른 대안을 제시합니다
• 비즈니스가 특이한 경우, 있는 그대로 설명하세요 — AI가 적절한 슬롯에 매핑해 드립니다

대화는 보통 10~20분 정도 소요됩니다.

---

3단계 — 팀 설계 문서 저장하기

AI가 최종 팀 설계 문서를 출력하면, 복사하여 접근하기 쉬운 곳에 저장해 두세요.

이 문서는 섹션 3에서 실제 에이전트 파일, 즉 각 에이전트가 작동하도록 하는 설정 파일을 생성할 때 사용합니다.