Skip to content

Latest commit

 

History

History
210 lines (150 loc) · 7.41 KB

GETTING_STARTED.md

File metadata and controls

210 lines (150 loc) · 7.41 KB

🚀 Getting Started with Simple CLI

Welcome to Simple CLI, the Universal AI Integration Platform. This guide will walk you through installing the tool, setting up your first "Company Context" (Digital Agency), and running your first autonomous tasks.

⚡ 0. First Steps (5 Minutes)

Before diving into configuration, take the Interactive Tour to see what Simple CLI can do.

Run the Onboarding Wizard:

simple onboard

This comprehensive, interactive wizard (First-Day Experience) will guide you through:

  1. Environment Check: Auto-detects dependencies and keys.
  2. Company Context Setup: Defines your Digital Agency's brand voice and goals.
  3. Framework Integration: Demonstrates the "Ingest-Digest-Deploy" cycle (e.g., Roo Code).
  4. SOP Execution: Creates and runs a real Standard Operating Procedure (initialize_project.sop).
  5. Ghost Mode: Activates the Job Delegator and schedules background tasks.
  6. HR Loop: Demonstrates the system's self-improvement capability.
  7. Operational Dashboard: Launches the real-time metrics view.

Read the Quick Start Guide for more details.

📦 1. Installation

Prerequisites

  • Node.js: Version 18.0.0 or higher.
  • Git: Installed and available in your PATH.
  • Docker (Optional): Recommended for "Ghost Mode" (24/7 background agents).

Install via NPM

Install the CLI globally to access the simple command from anywhere:

npm install -g @stan-chen/simple-cli

Configuration

Simple CLI requires an LLM provider to function. Create a .env file in your project root or export the variables in your shell profile (~/.zshrc or ~/.bashrc).

# Required: Primary Brain
export OPENAI_API_KEY="sk-..."

# Optional: Specialized Agents
export ANTHROPIC_API_KEY="sk-..." # For Claude
export GEMINI_API_KEY="..."       # For Google Gemini
export DEEPSEEK_API_KEY="..."     # For DeepSeek V3/R1

🏢 2. Setting Up Your Digital Agency

Simple CLI isn't just a chatbot; it simulates a team of digital employees. To start, you need to "Onboard" a company. This process sets up the Company Context, Brain, and SOPs.

Run the onboarding command:

simple onboard-company MyTechStartup

What just happened? (The 6-Pillar Setup)

The onboarding command automatically executed a Standard Operating Procedure (SOP) that established the 6 core pillars of your digital agency:

  1. Company Context: Created .agent/companies/MyTechStartup/config/company_context.json. This file tells the agent who it works for (Brand Voice, Tech Stack, Goals).
  2. The Brain: Initialized a Vector Database in .agent/companies/MyTechStartup/brain/ to remember past interactions and learnings.
  3. SOPs: Created a sops/ directory for your operational procedures.
  4. Ghost Mode: Scheduled background tasks (Job Delegator, Reviewer) in .agent/scheduler.json.
  5. HR Loop: Enabled the "Daily HR Review" task to analyze logs and propose self-improvements.
  6. Health Monitor: Activated the system health tracking.

Verify the Setup

Check that your company is active:

simple company status

You can list all available companies:

simple company list

⚡ 3. Running Your First Task

Now that your digital agency is set up, let's assign a task.

Interactive Mode

Just run simple and describe what you need. The agent will use its Smart Router to pick the best tool for the job.

simple "Create a 'Hello World' Python script and write a test for it."

What happens next:

  1. Planning: The agent analyzes the request.
  2. Delegation: It might delegate the coding to a sub-agent (e.g., coding_agent).
  3. Execution: It writes hello.py and test_hello.py.
  4. Verification: It runs the test to ensure it passes.

Direct Command

You can also run one-off commands without entering the interactive shell:

simple "Refactor src/utils.ts to be more readable"

🌟 4. The 4 Pillars in Practice

Understanding these concepts will help you get the most out of Simple CLI.

1. Company Context (The "Briefcase")

Every time the agent runs, it consults the Company Context.

  • Try it: Edit .agent/companies/MyTechStartup/config/company_context.json and change the "brand_voice" to "Pirate".
  • Run: simple "Write a commit message for the last change"
  • Result: "Arrr! We shipped a bounty of bug fixes!"

2. SOP-as-Code (The "Operating Manual")

Don't rely on prompts; rely on procedures.

  • Create: Write a Markdown file in sops/deployment.md.
  • Run: simple "Run the deployment SOP"
  • Result: The agent follows your checklist step-by-step, handling errors and retries automatically.

3. Ghost Mode (The "Night Shift")

Your agents work while you sleep.

  • Check: Look at .agent/scheduler.json. You'll see tasks like job-delegator scheduled to run hourly.
  • Logs: Check .agent/ghost_logs/ to see what your digital employees did overnight.

4. HR Loop (The "Self-Correction")

The system improves itself.

  • Mechanism: Every night, the HR agent analyzes execution logs.
  • Outcome: If a tool failed repeatedly, it might propose a code change to fix it. Check .agent/hr/proposals/ for pending updates.

🔌 5. Framework Integration (Ingest-Digest-Deploy)

Simple CLI can "absorb" other AI tools.

  • Ingest: Learn a new tool's CLI (e.g., aider).
  • Digest: Wrap it in an MCP Server (src/mcp_servers/aider).
  • Deploy: Register it in mcp.json.

Now, simple can use aider as a subordinate agent!

Power User Path

Want to build your own integration? Take the Rapid Framework Integration Tutorial to learn how to add "Roo Code" in 15 minutes.

🎭 6. Configuring Persona & Interfaces

Your digital agent can mimic human behavior (typing, working hours) across Slack, Teams, and Discord.

Configuration

Edit .agent/config/persona.json (or .agent/companies/<name>/config/persona.json):

{
  "name": "Sarah_DevOps",
  "role": "Senior DevOps Engineer",
  "voice": {
    "tone": "Professional but friendly"
  },
  "emoji_usage": true,
  "working_hours": "09:00-17:00",
  "response_latency": {
    "min": 500,
    "max": 2000
  },
  "catchphrases": {
    "greeting": ["Hey team!", "Hi all,"],
    "signoff": ["Best,", "Cheers!"]
  }
}

Interface Behavior

  • Working Hours: If you message the agent outside working_hours, it will reply with an "Out of Office" message and will not execute tasks.
  • Typing Indicators: The agent simulates typing latency (response_latency) and displays "Typing..." indicators in Slack/Teams/Discord while thinking.
  • Styling: Responses are automatically styled with the configured tone, emojis, and catchphrases.

📚 7. Advanced Guides

Ready for production? Check out our real-world deployment playbooks:

🎥 Next Steps

  • Try the Showcase: Run npm run demo to see a full 24-hour simulation in 2 minutes. Read more.
  • Deploy to Kubernetes: Ready for production? Read the Deployment Guide.
  • Customize: Edit persona.json to change your agent's personality.

Happy Building! 🚀