How mikeored/codex-of-agents works
High-Level Components
a) Client Layer (Top)
• React Frontend (planned)
• API Gateway Interface
b) Application Layer
• Use Cases (e.g., InvokeScribeUseCase)
• DTOs and Schemas
c) Domain Layer (Core)
• Entities (Document, Actor, Location, Agent)
• Domain Services
• Repositories Interfaces
d) Infrastructure Layer (Bottom)
• AWS Services (Bedrock, Lambda, DynamoDB)
• Repository Implementations
• Adapters (Primary/Secondary)
Component Relationships
a) Primary Flow:
Client Request → API Gateway → Primary Adapters → Use Cases → Domain Services → Secondary Adapters → AWS Services
b) Key Interactions:
• Agent interactions with LLM Services
• Chronicle execution flow
• Repository patterns for data persistence
Specific Architectural Elements
Create distinct sections for:
a) Agentic System:
• Agent Entities
• Tools System
• Chronicle Management
• Session Handling
b) Content Management:
• Document Hierarchy
• Actor/Location Management
• Content Relationships
Important Details
a) Domain Bounded Contexts:
• Document Management Context
• Agent Management Context
• Session Management Context
b) Cross-Cutting Concerns:
• Configuration Management
• Error Handling
• Schema Validation
Specific Implementation Notes
Key patterns:
• Hexagonal Architecture ports/adapters
• Repository Pattern implementations
• Event-driven interactions
• Tool system extensibility points
AWS Infrastructure Components
Highlight:
• API Gateway endpoints
• Lambda functions
• DynamoDB tables
• Bedrock integration points
Session management
Remember to:
• Clear boundaries between bounded contexts
• Hexagonal architecture's ports and adapters
• Demonstrate the flow of domain events
• Integration points with AWS services