The Connectivity Layer
Relevant source files
The following files were used as context for generating this wiki page:
The Connectivity Layer serves as the primary communication bridge between AI agents and the underlying infrastructure managed by Nexus Orchestrator. It functions as an MCP-based (Model Context Protocol) interface that enables autonomous agents to interact with environments, manage SSH connections, and access context files within a secure, local-first framework.
As one of the four core pillars of the system, this layer bridges the "infrastructure gap" by providing a standardized protocol for non-human identities to execute commands and manage state across distributed operating environments.
Sources: README.md:1-12, README.md:14-15
Architectural Overview
The Connectivity Layer is designed as an autonomous infrastructure daemon. It facilitates secure communication through two primary mechanisms: Model Context Protocol (MCP) for agent-to-orchestrator messaging and SSH for orchestrator-to-environment execution.
Communication Flow
The following diagram illustrates how an AI agent uses the Connectivity Layer to interact with managed infrastructure:
This flow ensures that agents do not require direct credentials for target environments, instead relying on the Nexus Orchestrator to broker requests through a Zero Trust security model.
Sources: README.md:6-15, context/AGENTS.md:1-5
Key Components
The Connectivity Layer is composed of several modules responsible for maintaining persistent state and establishing secure tunnels.
| Component | Description | Reference |
|---|---|---|
| MCP Server | Implements the Model Context Protocol to handle agent requests. | README.md:15 |
| SSH Manager | Handles connectivity to remote environments and infrastructure. | README.md:7 |
| Context Bridge | Manages the interaction between connectivity and the AGENTS.md file. | context/AGENTS.md:3-5 |
Model Context Protocol (MCP) Integration
The system uses MCP as the standardized protocol for agent-infrastructure communication. This allows agents to discover capabilities and execute tasks within the orchestration layer without needing custom integrations for different infrastructure providers.
Sources: README.md:15, nexus/connectivity/mcp_server.py
SSH and Environment Management
The Connectivity Layer autonomously manages SSH connectivity. This includes the lifecycle of connections required for agents to perform tasks in remote or containerized environments. By abstracting SSH management, the layer provides a seamless "nervous system" for the orchestrator.
Sources: README.md:7, nexus/connectivity/ssh_manager.py
Security and Trust
Connectivity is governed by a Zero Trust security model. Every agent interacting with the Connectivity Layer must be identified and assigned a trust level, which is tracked via the context system.
Identity and trust levels are persisted in the context files to ensure that the Connectivity Layer only processes authorized infrastructure changes.
| Field | Type | Description |
|---|---|---|
| Agent ID | String | Unique identifier for the interacting AI agent. |
| Trust Level | String | Defined security clearance for infrastructure access. |
Sources: README.md:18, context/AGENTS.md:7-10
Infrastructure Orchestration
The Connectivity Layer works in tandem with the Orchestration Layer (powered by LangGraph) to ensure that environment state is synchronized with the agent's intent. While the Orchestration Layer manages the workflow logic, the Connectivity Layer handles the physical execution and data transmission.
Sources: README.md:16, nexus/core/orchestrator.py
Summary
The Connectivity Layer is the essential "nervous system" of the Nexus Orchestrator. By combining MCP-based communication with autonomous SSH management and Zero Trust governance, it provides a robust framework for AI agents to safely and efficiently operate on local and remote infrastructure.