EagleEye

AI-powered unified search Slack bot using Claude Agent SDK and MCP for Slack, Notion, Linear, and GitHub.

Overview

EagleEye is a Slack bot that enables natural language search across your team's tools:

• Slack: Search messages and conversations
• Notion: Search pages and documents
• Linear: Search issues and projects
• GitHub: Search code, issues, and pull requests

Powered by Claude AI and the Model Context Protocol (MCP), EagleEye understands your questions and finds relevant information across all connected platforms.

Features

• Natural language search across multiple platforms
• Real-time progress updates during search
• Query-based server filtering for faster responses
• Slack Block Kit formatted responses
• GitHub organization filtering support

Architecture

┌─────────────────────────────────────────────────────────────┐
│                     Slack Workspace                          │
│                    ┌─────────────┐                           │
│                    │ @EagleEye   │                           │
│                    └──────┬──────┘                           │
└───────────────────────────┼──────────────────────────────────┘
                            │
                            ▼
┌─────────────────────────────────────────────────────────────┐
│                    EagleEye Bot (Python)                     │
│  ┌─────────────────────────────────────────────────────┐   │
│  │              Slack Bolt (Socket Mode)                │   │
│  └─────────────────────────┬───────────────────────────┘   │
│                            │                                 │
│  ┌─────────────────────────▼───────────────────────────┐   │
│  │           Claude Agent SDK + MCP Servers             │   │
│  │  ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐   │   │
│  │  │  Slack  │ │ Notion  │ │ Linear  │ │ GitHub  │   │   │
│  │  └─────────┘ └─────────┘ └─────────┘ └─────────┘   │   │
│  └─────────────────────────────────────────────────────┘   │
└─────────────────────────────────────────────────────────────┘

Requirements

• Python 3.12+
• Node.js 20+ (for MCP servers via npx)
• uv (recommended) or pip
• Claude CLI (@anthropic-ai/claude-code)
• Slack workspace with bot permissions

Quick Start

1. Clone and setup

git clone https://github.com/your-org/eagleeye.git
cd eagleeye

# Using uv (recommended)
uv sync

# Or using pip
pip install -e ".[dev]"

2. Configure environment

cp .env.example .env
# Edit .env with your credentials

3. Claude CLI authentication

npm install -g @anthropic-ai/claude-code
claude login

4. Run the bot

uv run python -m src

# With debug logging
DEBUG=true uv run python -m src

Slack App Setup

1. Go to Slack API and create a new app
2. Enable Socket Mode and get the App-Level Token (xapp-...)
3. Add Bot Token Scopes:
   • app_mentions:read
   • chat:write
   • channels:history
   • channels:read
4. Install the app to your workspace and get the Bot Token (xoxb-...)

API Keys

Service	Where to get
Slack Bot Token	Slack App > OAuth & Permissions
Slack App Token	Slack App > Basic Information > App-Level Tokens
Slack Team ID	Browser URL when logged into Slack
Notion API Key	Notion Integrations
Linear API Key	Linear Settings > API
GitHub Token	GitHub Settings > Tokens

Usage

Mention the bot in any channel:

@EagleEye 지난주 배포 관련 논의
@EagleEye authentication bug in Linear
@EagleEye ai-tutor 레포 최근 PR

EagleEye will search the relevant platforms and return formatted results with links.

Configuration

Environment Variable	Required	Description
SLACK_BOT_TOKEN	Yes	Slack bot OAuth token
SLACK_APP_TOKEN	Yes	Slack app-level token for Socket Mode
SLACK_SIGNING_SECRET	Yes	Slack app signing secret
SLACK_TEAM_ID	Yes	Slack workspace ID
NOTION_API_KEY	Yes	Notion integration API key
LINEAR_API_KEY	Yes	Linear API key
GITHUB_TOKEN	No	GitHub personal access token
GITHUB_ORG	No	Limit GitHub search to specific organization
CLAUDE_MODEL	No	Claude model (default: claude-sonnet-4-20250514)
DEBUG	No	Enable debug logging (default: false)
ENABLE_SLACK_MCP	No	Enable Slack search (default: true)
ENABLE_NOTION_MCP	No	Enable Notion search (default: true)
ENABLE_LINEAR_MCP	No	Enable Linear search (default: true)
ENABLE_GITHUB_MCP	No	Enable GitHub search (default: true)

Deployment

Docker (Recommended)

# 1. Authenticate Claude CLI locally
claude login

# 2. Configure environment
cp .env.example .env
# Edit .env with your API keys

# 3. Build and run
docker compose up -d

# View logs
docker compose logs -f

Manual Deployment

# Install dependencies
sudo yum install -y docker nodejs  # Amazon Linux
# or
sudo apt install -y docker.io nodejs  # Ubuntu

# Install Claude CLI
sudo npm install -g @anthropic-ai/claude-code
claude login

# Clone and configure
git clone <repo-url>
cd eagleeye
cp .env.example .env
# Edit .env

# Run with Docker
docker build -t eagleeye .
docker run -d \
  --name eagleeye \
  --restart unless-stopped \
  --env-file .env \
  -v ~/.claude:/root/.claude:ro \
  eagleeye

Development

Run tests

uv run pytest tests/ -v

Code quality

# Linting
uv run ruff check src/ tests/

# Type checking
uv run mypy src/

# Format
uv run ruff format src/ tests/

Project structure

eagleeye/
├── src/
│   ├── __main__.py         # Entry point
│   ├── app.py              # Slack bot application
│   ├── claude_agent.py     # Claude Agent SDK integration
│   ├── config.py           # Settings management
│   ├── log_config.py       # Structured logging
│   ├── slack_formatter.py  # Slack Block Kit formatting
│   └── models/
│       └── search.py       # Search result models
├── tests/
│   ├── test_app.py
│   ├── test_claude_agent.py
│   ├── test_models.py
│   └── test_slack_formatter.py
├── Dockerfile
├── docker-compose.yml
├── pyproject.toml
├── .env.example
└── README.md

How it works

1. User mentions @EagleEye with a search query
2. EagleEye detects relevant platforms from the query keywords
3. Claude Agent SDK processes the query with filtered MCP servers
4. MCP servers search Slack, Notion, Linear, and GitHub
5. Claude synthesizes results into a Slack mrkdwn formatted response
6. Real-time progress updates are shown during the search
7. Final response is sent with Block Kit formatting and source links

License

MIT