graph LR
Application_Core_MCPApp_["Application Core (MCPApp)"]
Configuration_Management_Settings_["Configuration Management (Settings)"]
Decorator_Registry["Decorator Registry"]
Signal_Registry["Signal Registry"]
Task_Registry_ActivityRegistry_["Task Registry (ActivityRegistry)"]
MCP_Server_Management["MCP Server Management"]
LLM_Selection["LLM Selection"]
Application_Core_MCPApp_ -- "uses" --> Configuration_Management_Settings_
Application_Core_MCPApp_ -- "manages" --> Decorator_Registry
Application_Core_MCPApp_ -- "manages" --> Signal_Registry
Application_Core_MCPApp_ -- "manages" --> Task_Registry_ActivityRegistry_
Application_Core_MCPApp_ -- "orchestrates" --> MCP_Server_Management
Application_Core_MCPApp_ -- "integrates" --> LLM_Selection
Configuration_Management_Settings_ -- "configures" --> LLM_Selection
Configuration_Management_Settings_ -- "configures" --> MCP_Server_Management
MCP_Server_Management -- "depends on" --> Configuration_Management_Settings_
LLM_Selection -- "depends on" --> Configuration_Management_Settings_
This component is the foundational layer of the mcp-agent application. It is responsible for the initial bootstrapping of the entire agent framework, establishing the global application context, and managing the loading, parsing, and provision of all application settings and sensitive information. It acts as the central orchestrator for the agent's environment, ensuring that all subsequent components operate with the correct parameters and have access to essential services. Its design emphasizes modularity and extensibility, crucial for an AI agent development framework.
The primary entry point and orchestrator of the mcp-agent application. It initializes and coordinates all other core services and components, setting up the runtime environment for agents and workflows.
Related Classes/Methods:
Centralized management for all application configurations, including API keys, service endpoints, and operational parameters. It ensures that all components have access to the necessary settings in a structured and secure manner, supporting various LLM providers and external services.
Related Classes/Methods:
mcp_agent.config.Settingsmcp_agent.config.AnthropicSettingsmcp_agent.config.OpenAISettingsmcp_agent.config.TemporalSettingsmcp_agent.config.AzureSettingsmcp_agent.config.BedrockSettingsmcp_agent.config.GoogleSettings
A central registry for managing and retrieving decorators, enabling dynamic extension and modification of component behavior across the framework. This supports the "Strategy Pattern" and extensibility by allowing new functionalities to be plugged in without altering core code.
Related Classes/Methods:
Provides a mechanism for registering and dispatching signals, facilitating event-driven communication and coordination between different parts of the agent framework. This is crucial for implementing the "Observer Pattern" and asynchronous interactions.
Related Classes/Methods:
A registry for managing and retrieving executable tasks or activities, allowing the framework to dynamically discover and invoke operations. This is key for defining and executing agent capabilities.
Related Classes/Methods:
Manages the lifecycle and connections to various Model Context Protocol (MCP) servers, enabling the agent to interact with external models and services. This component is critical for the "Microservices/Service-Oriented Architecture" aspect of the framework.
Related Classes/Methods:
mcp_agent.mcp.mcp_connection_manager.MCPConnectionManagermcp_agent.mcp.mcp_server_registry.ServerRegistrymcp_agent.mcp.mcp_aggregator.MCPAggregator(77:1357)
Responsible for selecting and managing Large Language Models (LLMs), including their configuration and augmentation, to provide the core AI capabilities for the agents. It allows for flexible integration of different LLM providers.
Related Classes/Methods: