Lebot - AI-Powered Slack Assistant

A sophisticated Slack Assistant Bot that integrates with Anthropic's Claude API to provide intelligent AI assistance directly in Slack workspaces. Built with Python, Slack Bolt, and async architecture for high performance.

Features

• 🤖 Slack Assistant Integration: Native Slack Assistant API for seamless conversational interfaces
• 🧠 Claude AI Models:
  • Normal Mode: Claude Sonnet 4 (optimized for speed)
  • Beast Mode: Claude Opus 4 (maximum intelligence) - selectable via interactive buttons
• 🔧 Tool Integration: Extensible architecture with built-in bash commands and Linear API integration
• ⚡ Real-time Streaming: Async streaming responses with live status updates
• 💭 Thinking Mode: Shows Claude's reasoning process with proper formatting
• 🛑 Emergency Stop: Cancel long-running operations with stop button
• 📊 Linear Integration: Query and manage Linear issues directly from Slack

Quick Start

Prerequisites

• Python 3.8+
• Slack workspace with admin permissions
• Anthropic API key
• Linear API key (optional)

Installation

1. Clone the repository

git clone https://github.com/yourusername/lebot.git
cd lebot

2. Set up Python environment

python3 -m venv .venv
source .venv/bin/activate  # Windows: .\.venv\Scripts\Activate
pip install -r requirements.txt

3. Create Slack App

   • Go to api.slack.com/apps/new
   • Choose "From an app manifest"
   • Copy contents of manifest.json and paste in the JSON tab
   • Click Create, then Install to Workspace

4. Configure environment variables

cp .env.example .env

Edit .env with your credentials:

SLACK_BOT_TOKEN=xoxb-your-bot-token        # From OAuth & Permissions
SLACK_APP_TOKEN=xapp-your-app-token        # From Basic Information > App-Level Tokens
ANTHROPIC_API_KEY=sk-ant-api03-your-key    # From Anthropic Console
LINEAR_OAUTH_KEY=lin_api_your_key          # Optional: From Linear Settings

5. Run the bot

Using Slack CLI (recommended):

# Install Slack CLI if you haven't already
# Visit: https://api.slack.com/automation/cli/install

# Run the app with Slack CLI
slack run

Or run directly with Python:

python3 app.py

Usage

1. Start a conversation: Click the ✨ Assistant button in any Slack channel or DM to start a new assistant thread
2. Select AI model: Use the interactive buttons in the greeting:
   • 🚀 Normal Mode: Fast responses with Claude Sonnet 4
   • 🦾 Beast Mode: Maximum capability with Claude Opus 4
3. Ask questions: Type naturally in the assistant thread - no @ mentions needed
4. Use tools: The bot can execute bash commands and query Linear issues

Example Interactions

In an assistant thread, simply type:

Can you help me understand this Python error?
Show me all in-progress Linear issues
What issues were created this week?
Can you analyze our test coverage?

Architecture

User ↔ Slack Assistant ↔ LLM (Claude) ↔ Tools

Key Components

• app.py: Main entry point with async Slack Bolt app
• slack_hook/: Core Slack integration
  • assistant.py: Async event handlers with decorators
  • claude.py: Custom async Claude client with streaming
  • conversation_manager.py: Thread context management
  • message_parser.py: Parsing for thinking blocks and tool uses
• tools/: Extensible tool system
  • graphql.py: GraphQL client with Linear API integration

Async Architecture

The entire application uses async/await for optimal performance:

• Non-blocking I/O operations
• Real-time streaming responses
• Concurrent tool execution
• Proper resource management

Development

Code Quality

# Linting
flake8 *.py && flake8 slack_hook/ && flake8 tools/

# Formatting (125 char line length)
black .

# Testing
pytest

Linear API Usage

from tools.graphql import LinearClient

# Initialize client
linear = LinearClient(os.getenv('LINEAR_OAUTH_KEY'))

# Get issues
issues = linear.get_issues(limit=50)
in_progress = linear.get_in_progress_issues()

# Query by date
from datetime import datetime, timedelta
weekly_issues = linear.get_issues_by_date_range(
    start_date=(datetime.now() - timedelta(days=7)).strftime('%Y-%m-%d'),
    end_date=datetime.now().strftime('%Y-%m-%d')
)

Configuration

Slack App Manifest

The manifest.json defines:

• Assistant feature with custom description
• Required OAuth scopes
• Event subscriptions
• Socket Mode configuration

Environment Variables

See .env.example for all available options:

• SLACK_BOT_TOKEN: Bot user OAuth token
• SLACK_APP_TOKEN: App-level token for Socket Mode
• ANTHROPIC_API_KEY: Claude API key
• LINEAR_OAUTH_KEY: Linear API key (optional)

Future Enhancements

See docs/claude-code-integration-architecture.md for planned Claude Code integration that would add:

• File operations and code generation
• Dynamic script discovery and execution
• Sandboxed development environment
• Enhanced tool capabilities

Troubleshooting

Common Issues

1. Bot not responding: Check Socket Mode connection and app tokens
2. Linear integration failing: Verify API key and permissions
3. Streaming issues: Ensure async handlers are properly awaited
4. Type errors: Run flake8 to check for type hint issues

Debug Mode

Enable debug logging by setting:

export DEBUG=true

Contributing

1. Fork the repository
2. Create a feature branch
3. Run tests and linting
4. Submit a pull request

License

MIT License - see LICENSE file for details

Support

• Slack API Documentation
• Anthropic Claude Documentation
• Linear API Documentation

Security

• Store all API keys in .env file
• Never commit credentials to version control
• Use environment variables for all sensitive data
• Review manifest.json OAuth scopes before deployment