a2a-mcp-contractor-automation

# A2A MCP Contractor Email Automation A lightweight demo application showcasing Agent-to-Agent (A2A) communication for automating email management in a small business contractor setting. ## Features - Automated email processing and classification - Intelligent response generation for quote requests - Daily email interaction summaries - Clean A2A architecture with three collaborating agents - Local AI infrastructure integration - Containerized deployment ## Architecture The system consists of three main agents: 1. **EmailProcessorAgent** - Monitors and processes incoming emails - Classifies email intent using local LLM - Delegates tasks to appropriate agents - Manages artifact exchange 2. **ResponseAgent** - Generates customized responses for quote requests - Uses local templates and LLM for response customization - Maintains response consistency 3. **SummaryAgent** - Creates concise summaries of important email interactions - Maintains daily logs of email summaries - Uses local LLM for summarization ## Prerequisites 1. Docker and Docker Compose 2. Python 3.13 3. Local LLM (Ollama or LM Studio) 4. ChromaDB 5. Gmail API credentials ## Setup Instructions 1. **Install Dependencies** ```bash pip install -r requirements.txt ``` 2. **Set up Ollama** ```bash # Install Ollama curl https://ollama.ai/install.sh | sh # Run Ollama ollama serve # Pull a model (e.g., llama2) ollama pull llama2 ``` 3. **Set up Gmail API** - Go to Google Cloud Console - Create a new project - Enable Gmail API - Create OAuth 2.0 credentials - Download credentials.json and place it in the project root 4. **Environment Variables** Create a `.env` file with: ``` GMAIL_API_CREDENTIALS=/path/to/credentials.json LLM_URL=http://localhost:11434/api/generate ``` 5. **Run with Docker** ```bash # Build and run containers docker-compose up --build ``` 6. **Run Directly** ```bash # Run EmailProcessorAgent python agents/email_processor_agent.py # Run ResponseAgent python agents/response_agent.py # Run SummaryAgent python agents/summary_agent.py ``` ## Usage 1. **Process Emails** ```bash curl -X POST http://localhost:8001/process_email \ -H "Content-Type: application/json" \ -d '{ "sender": "customer@example.com", "subject": "Request for Quote", "body": "I'm interested in your services..." }' ``` 2. **Check Outputs** - Responses: `output/templates/` - Summaries: `output/summaries/daily_summary_YYYY-MM-DD.txt` ## Code Structure ``` a2a-mcp-contractor-automation/ ├── agents/ │ ├── email_processor_agent.py │ ├── response_agent.py │ └── summary_agent.py ├── config.py ├── requirements.txt ├── Dockerfile ├── docker-compose.yml └── output/ ├── templates/ └── summaries/ ``` ## AI Integration The system uses: - Local LLM (Ollama/LM Studio) for: - Email intent classification - Response generation - Summary creation - ChromaDB for: - Context storage - Template management - Summary history ## License MIT License - see LICENSE file for details