One Thing to Understand Before You Start

In BALIA OS, you are the CEO.

OpenClaw has a default agent called main. You don't touch it, configure it, or use it. It stays in the background. You are the one giving orders.

This means the structure is simple:

INFO

You (CEO) ↓ direct commands via VS Code or OpenGoat UI ├── Agent #1 (ceo — your digital twin) → reports to #agent1-channel in Discord ├── Agent #2 (free slot) → reports to #agent2-channel in Discord ├── Agent #3 (free slot) → reports to #agent3-channel in Discord ├── Agent #4 (free slot) → reports to #agent4-channel in Discord ├── Agent #5 (free slot) → reports to #agent5-channel in Discord └── Agent #6 (security-guardian — fixed) → reports to #security-channel in Discord

No middle manager. No orchestrator agent between you and your team. You speak directly to each agent. Each agent reports directly to their own Discord channel. You read the reports.

This is the design that works. Simple, direct, and no subagent errors.

---

Your 6 Slots — The Confirmed Layout

Before you design anything, here is the fixed frame your team fits into:

Slot	Role	Agent ID	Notes
#1	CEO (main agent — your digital twin)	ceo (fixed)	Holds the identity/ folder (Section 4). Your .env uses AGENT_ID=ceo (Section 3).
#2–#5	Free slots — designed for your business	You choose	Lowercase, no spaces (e.g. marcus, nova)
#6	Security	security-guardian (fixed)	The name you choose in this section is used only as the Discord channel display name.

Two agent IDs are fixed and must never be changed:

• Slot #1 is the CEO agent. This is your main agent — your digital twin, the one you talk to most. Its agent ID is always ceo. In Section 3 you'll set AGENT_ID=ceo in your .env, and in Section 4 it receives the identity/ folder.
• Slot #6 is the Security agent. Its agent ID is always security-guardian. You'll still pick a display name for it in this section — that name is used only for its Discord channel.

Slots #2 through #5 are yours to design freely with the prompt below.

---

How This Section Works

Copy the prompt below and paste it into your chat AI. The AI will guide you through designing your team step by step — asking one question at a time, helping you think through decisions, and building your team profile as you go.

By the end of the conversation, the AI will produce a completed Team Design Document. Save it. You'll bring it into Section 3 to create your agent files.

If you're not sure about something, just say so. The AI will help you figure it out.

---

▶ Paste this into your AI to get started

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:
- Slot #1: CEO (main agent — my digital twin). Agent ID: ceo (fixed — do not change).
- Slots #2 through #5: I configure these for my business.
- Slot #6: Always Security — agent ID: security-guardian (fixed). I will choose
  a display name for it, but the agent ID 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 4 roles (Slot #2 through #5).
(Slot #1 is my CEO agent — already fixed. Slot #6 is Security — already fixed.)
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 — CEO (MAIN AGENT)
Agent ID: ceo (fixed — do not change)
Name:
Role title:
Mission:
Personality:
Language:
Discord channel:

SLOT #2 — [ROLE NAME]
Agent ID:
Name:
Role title:
Mission:
Personality:
Language:
Discord channel:

SLOT #3 — [ROLE NAME]
Agent ID:
Name:
Role title:
Mission:
Personality:
Language:
Discord channel:

SLOT #4 — [ROLE NAME]
Agent ID:
Name:
Role title:
Mission:
Personality:
Language:
Discord channel:

SLOT #5 — [ROLE NAME]
Agent ID:
Name:
Role title:
Mission:
Personality:
Language:
Discord channel:

SLOT #6 — SECURITY
Agent ID: security-guardian (fixed — do not change)
Display name:
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.

---

Step 2 — Have the Conversation

Work through the questions with your AI concierge. A few tips:

• If you're not sure what roles you need, describe what you want to achieve — the AI will suggest the structure
• If a suggested name doesn't feel right, say so — the AI will offer alternatives
• If your business is unusual, just describe it plainly — the AI will map it to the right slots

The conversation usually takes 10–20 minutes.

---

Step 3 — Save Your Team Design Document

If you haven't completed Section 3 yet, you won't be able to save it with VS Code + Claude Code yet. In that case, copy the AI's Team Design Document output into a text file for safekeeping for now, and save it to $env:USERPROFILE\.opengoat\identity\team_design.md (Mac: ~/.opengoat/identity/team_design.md) once you've completed Section 3.

When the AI produces the final Team Design Document, use VS Code + Claude Code to save it to the right location.

Save with VS Code + Claude Code (recommended):

Paste this into Claude Code in VS Code:

I have my team design document ready. Please save it to this path:

C:\Users\[your username]\.opengoat\identity\team_design.md

(On Mac: ~/.opengoat/identity/team_design.md)

Create the folder if it doesn't exist.

Content: [paste your team design document here]

Claude Code will automatically create the folder and save the file.

Save location:

$env:USERPROFILE\.opengoat\
└── identity\
    └── team_design.md   ← save here

From Section 3 onward, instead of attaching the team design document, you'll reference this file directly in VS Code.

Bring this document into Section 3. In Section 3, you'll use it to generate the actual configuration files that make each agent work.