Agent Playbook CLI 🤖

Transform your ideas into comprehensive AI agent playbooks using AI-guided discovery.

✨ Features

• AI-Powered Discovery: Leverages Claude or Gemini AI to ask targeted questions about your agent concept
• Structured Workflow: Three rounds of increasingly specific questions to refine your vision
• Quick Number Selection: Press 1-9 to instantly select answers, 0 for custom input, 's' to skip
• Agent Playbooks: Generates comprehensive playbooks including behavior patterns, capabilities, and implementation details
• File Structure Generation: Creates an optimal project structure with all necessary files and directories
• Session Management: Save and resume sessions, never lose your progress
• Beautiful CLI: Intuitive interface with colors, progress indicators, and clear navigation
• Export Options: Export to Markdown or JSON for easy sharing and processing

🎯 How It Works

1. Initial Idea: Start with your agent concept
2. Discovery Questions: Answer 3 rounds of AI-generated multiple-choice questions
   • Round 1: General requirements and audience
   • Round 2: User experience and differentiation
   • Round 3: Technical implementation details
3. Agent Playbook: AI generates a comprehensive agent playbook
4. File Structure: Creates an optimal project structure
5. Export: Save as Markdown or JSON for your team

📦 Installation

npm install -g agent-playbook

# Or run directly with npx
npx agent-playbook

🚀 Quick Start

# Start interactive mode
agent-playbook

# Start with an idea
agent-playbook new --idea "Customer service AI agent"

# Resume a session
agent-playbook resume

# List all sessions
agent-playbook list

📖 Commands

agent-playbook new

Start a new agent playbook session.

Options:

• -i, --idea <idea> - Start with a predefined idea
• -s, --skip-questions - Use default configuration

agent-playbook resume

Resume a previous session from where you left off.

agent-playbook list

Display all your sessions with their status and last update time.

agent-playbook export [sessionId]

Export a session to markdown or JSON format.

Options:

• -f, --format <format> - Export format: markdown, json, or both

agent-playbook clean

Remove old sessions to free up space.

Options:

• -d, --days <days> - Keep sessions newer than this many days (default: 30)
• -f, --force - Skip confirmation prompt

agent-playbook config [action]

Manage configuration and API keys.

Actions:

• reset - Reset all settings to defaults
• keys - Manage API keys

⚙️ Configuration

API Keys

Set your API keys as environment variables:

export ANTHROPIC_API_KEY=your_key_here
export GEMINI_API_KEY=your_key_here

Or let the CLI prompt you and save them securely.

Session Options

During setup, you can configure:

• Number of questions per round (2-8 for first round, 2-6 for subsequent)
• Answer choices per question (2-6 options)
• AI provider (Anthropic Claude or Google Gemini)

📁 Output Example

Generated Agent Playbook

# Customer Service AI Agent

## Executive Summary
An intelligent customer service agent that handles inquiries, provides support, and escalates complex issues...

## Core Capabilities
1. Natural Language Understanding
2. Multi-turn Conversations
3. Issue Resolution
...

Generated File Structure

customer-service-agent/
├── src/
│   ├── agents/
│   │   ├── CustomerServiceAgent.ts
│   │   ├── ConversationHandler.ts
│   │   └── IssueResolver.ts
│   ├── services/
│   │   ├── ai-service.ts
│   │   └── knowledge-base.ts
│   └── handlers/
│       ├── index.ts
│       └── escalation.ts
├── package.json
└── README.md

🔒 Privacy & Security

• API keys are stored locally in your system's config directory
• Sessions are saved locally in ~/.agent-playbook/sessions/
• No data is sent to external servers except AI API calls
• You can clear all data with agent-playbook config reset

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📄 License

MIT License - see LICENSE file for details.