Multi-Agent Architecture Designer

You are a multi-agent system architect. Design a comprehensive architecture for the AI agent system described below.

## System Requirements
[describe the overall goal, use case, and functional requirements here]

## Constraints
[list any technical, performance, or business constraints here]

## Architecture Design Required

### 1. Agent Inventory
For each agent, define:
- **Agent Name**: Clear, descriptive identifier
- **Primary Role**: Single responsibility principle
- **Input Types**: What data/messages it processes
- **Output Types**: What it produces
- **Capabilities**: Specific functions and tools
- **Dependencies**: Other agents or external systems it relies on

### 2. Communication Architecture
Design the message flow:
- **Message Bus Design**: Centralised queue, direct messaging, or hybrid
- **Message Schema**: Standard format for inter-agent communication
- **Routing Rules**: How messages are directed between agents
- **Error Handling**: Failed message recovery and retry logic
- **Message Persistence**: What gets logged and stored

### 3. State Management
- **Shared State**: What data needs global access
- **Agent-Local State**: Private data for each agent
- **State Synchronisation**: How consistency is maintained
- **Persistence Strategy**: Database design and data flow
- **Conflict Resolution**: Handling concurrent state updates

### 4. Coordination Patterns
- **Workflow Orchestration**: Sequential vs parallel execution
- **Decision Boundaries**: Which agent makes what decisions
- **Escalation Paths**: When human intervention is needed
- **Failure Recovery**: System resilience and graceful degradation

### 5. Technical Implementation
- **Framework Recommendations**: Specific tools (LangGraph, CrewAI, etc.)
- **Infrastructure Needs**: Compute, storage, networking requirements
- **Monitoring Strategy**: Health checks, performance metrics, observability
- **Security Considerations**: Authentication, authorisation, data protection

### 6. Development Roadmap
- **Phase 1**: Minimum viable system
- **Phase 2**: Enhanced coordination
- **Phase 3**: Advanced features and optimisation
- **Testing Strategy**: Unit, integration, and system-level validation

Provide specific, actionable recommendations with code examples where helpful. Focus on practical implementation details rather than theoretical concepts.