Quick Start Guide
Get MCP Task Tracker up and running in 5 minutes!
Prerequisites
- Docker and Docker Compose (recommended)
OR
- Go 1.21+
- PostgreSQL 15+
Option 1: Docker (Recommended)
1. Clone and Start
git clone https://github.com/techbuzzz/agent-shaker.git
cd agent-shaker
docker-compose up -d
2. Verify
curl http://localhost:8080/health
Expected response: OK
3. Open Web UI
Open your browser: http://localhost:8080
4. Run Demo
./demo.sh
This creates:
- 1 project
- 2 agents (backend, frontend)
- 2 tasks
- 1 documentation entry
5. Stop Services
docker-compose down
Option 2: Local Development
1. Clone Repository
git clone https://github.com/techbuzzz/agent-shaker.git
cd agent-shaker
2. Install Dependencies
go mod download
3. Start PostgreSQL
# Using Docker
docker run -d \
--name mcp-postgres \
-e POSTGRES_DB=mcp_tracker \
-e POSTGRES_USER=mcp \
-e POSTGRES_PASSWORD=secret \
-p 5432:5432 \
postgres:15-alpine
# OR install PostgreSQL locally and create database
createdb mcp_tracker
4. Set Environment
export DATABASE_URL="postgres://mcp:secret@localhost:5432/mcp_tracker?sslmode=disable"
5. Build and Run
make build
./mcp-server
Or:
make run
6. Verify
curl http://localhost:8080/health
First Steps
Create a Project
curl -X POST http://localhost:8080/api/projects \
-H "Content-Type: application/json" \
-d '{
"name": "My First Project",
"description": "Testing MCP Task Tracker"
}'
Save the id from the response (you'll need it for next steps).
Register an Agent
curl -X POST http://localhost:8080/api/agents \
-H "Content-Type: application/json" \
-d '{
"project_id": "YOUR-PROJECT-ID",
"name": "My Agent",
"role": "backend",
"team": "Development Team"
}'
Save the agent id.
Create a Task
curl -X POST http://localhost:8080/api/tasks \
-H "Content-Type: application/json" \
-d '{
"project_id": "YOUR-PROJECT-ID",
"title": "My First Task",
"description": "Testing task creation",
"priority": "high",
"created_by": "YOUR-AGENT-ID",
"assigned_to": "YOUR-AGENT-ID"
}'
View in Web UI
- Open http://localhost:8080
- You should see your project
- Click on it to select
- Switch to "Agents" tab to see your agent
- Switch to "Tasks" tab to see your task
Using Makefile
# Build
make build
# Run locally
make run
# Run tests
make test
# Start with Docker
make docker-up
# Stop Docker services
make docker-down
# View logs
make docker-logs
# Run demo
make demo
# Clean build artifacts
make clean
# Show all commands
make help
Next Steps
-
Read the Documentation
- README.md - Overview and features
- docs/API.md - Complete API reference
- docs/COPILOT_INTEGRATION.md - Copilot setup
-
Try the Demo
./demo.sh -
Integrate with GitHub Copilot
-
Explore the Web UI
- Create projects
- Register agents
- Create and manage tasks
- Add documentation
-
Test WebSocket Updates
- Open Web UI in two browser tabs
- Create a task in one tab
- Watch it appear in the other tab in real-time
Common Issues
Port Already in Use
If port 8080 is already in use:
# Change port in docker-compose.yml or set environment variable
export PORT=8081
./mcp-server
Database Connection Failed
Check that PostgreSQL is running:
# Docker
docker ps | grep postgres
# Local
pg_isready
WebSocket Not Connecting
Ensure you're using the correct project ID in the WebSocket URL:
ws://localhost:8080/ws?project_id=YOUR-PROJECT-ID
Getting Help
- Check docs/API.md for API details
- See CONTRIBUTING.md for development info
- Open an issue on GitHub for bugs or questions
What's Next?
- Multi-team Coordination: Set up multiple agents for different teams
- Documentation Hub: Add markdown documentation with tags
- Real-time Collaboration: Use WebSocket for live updates
- GitHub Copilot Integration: Connect your AI agents
Happy tracking! 🚀