Slack AI Agent Starter

A comprehensive toolkit for building and running AI-powered Slack bots using LangGraph and Slack Bolt API. This implementation integrates advanced AI capabilities through LangGraph and LangChain, supporting both OpenAI and Anthropic models.

Features

1. Conversation Capabilities

• Natural language interaction through mentions
• Context-aware responses using conversation history
• Thread-based communication management
• Interactive message components
• Support for both OpenAI and Anthropic models

2. Command System

• help - Display detailed usage instructions
• ai [question] - Direct interaction with AI agent
• hello - Interactive greeting with button interface

3. Core Functionalities

• Advanced natural language processing with LangGraph and LangChain
• Conversation history consideration
• Interactive UI elements (buttons, etc.)
• Comprehensive error handling and feedback
• Customizable workflows
• Tool execution and complex interactions

Prerequisites

• Python 3.11 or higher
• Poetry for dependency management
• Slack App credentials
• OpenAI API key (or Anthropic API key if using Claude)
• LangGraph setup

Setup

Environment Variables

Create a .env file with the following:

# Slack Configuration
SLACK_BOT_TOKEN=your-bot-token
SLACK_APP_TOKEN=your-app-token
SLACK_SIGNING_SECRET=your-signing-secret
PORT=3000  # Optional, defaults to 3000

# AI Configuration
OPENAI_API_KEY=your-openai-api-key
# Or for Anthropic:
# ANTHROPIC_API_KEY=your-anthropic-api-key

# LangGraph Configuration
LANGGRAPH_URL=your-langgraph-url
LANGGRAPH_TOKEN=your-langgraph-token

Installation

1. Clone the repository

git clone https://github.com/kyaukyuai/slack-ai-agent-starter.git
cd slack-ai-agent-starter

2. Install dependencies

poetry install

3. Start the development servers

# Start both LangGraph and Web servers
make dev

# Or start them separately
make dev-langgraph  # For LangGraph server
make dev-web        # For Web server

Usage

Basic Interaction

• Mention the bot: @AI Assistant hello
• Use direct commands: ai what's the weather?
• Get help: help

Best Practices

1. Use threads for related conversations
2. Be specific with questions
3. Provide context when needed
4. Use appropriate commands for different tasks

Project Structure

slack_ai_agent/
├── agents/                # AI agent implementation
│   ├── simple_agent.py   # LangGraph AI agent implementation
│   ├── security/         # Security-related functionality
│   └── utils/
│       ├── nodes.py      # Agent workflow nodes
│       ├── tools.py      # Tool implementations
│       └── types.py      # Type definitions
├── slack/
│   ├── handler/          # Event, message, and action handlers
│   ├── app.py           # Main Slack app configuration
│   └── utils.py         # Utility functions
└── README.md

Development

Key Dependencies

• slack-bolt: For Slack bot functionality
• langgraph: For AI agent workflow management
• langchain_core: Core LangChain functionality
• langchain_openai: OpenAI model integration
• langchain_anthropic: Anthropic model integration
• python-dotenv: For environment variable management

Development Tools

• ruff: For code linting
• mypy: For type checking
• pre-commit: For git hooks
• langgraph-cli: For development tools

Development Principles

• Type hints and docstrings for all functions and classes
• Modular design with separate packages for Slack bot and AI agent
• Clear separation of concerns between bot handling and AI processing
• Comprehensive error handling and logging

Error Handling

The implementation includes:

• Graceful handling of API errors
• Detailed logging for debugging
• User-friendly error messages in Slack
• Tool execution result reporting

Extending the Agent

To add new capabilities:

1. Define new tool functions in agents/utils/nodes.py
2. Update the agent workflow in agents/simple_agent.py
3. Add any new environment variables to .env
4. Update type definitions in agents/utils/types.py if needed

Contributing

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

License

This project is licensed under the MIT License - see the LICENSE file for details.