Overview

Synedrix OS is standards-aware at the interoperability layer. An agentic operating system needs a clear strategy for how agents connect to tools, context providers, and other agents beyond the platform boundary. This document defines the high-level interoperability and protocol strategy for Synedrix OS.

Strategic Position

Synedrix OS treats MCP as the preferred protocol family for agent-to-tool and agent-to-context integration, and A2A as the preferred protocol family for agent-to-agent interoperability across system boundaries. These are interoperability standards, not replacements for internal orchestration, governance, memory, or observability.

Protocol Roles

ProtocolPrimary roleScope in Synedrix OSWhat it is not
MCP (Model Context Protocol)Standardized access to tools, resources, prompts, and contextAgent-to-tool, agent-to-resource, copilot-to-system integrationNot a workflow engine, not a database, not a replacement for internal adapters
A2A (Agent2Agent)Standardized communication between autonomous agents and agent systemsCross-system agent delegation, partner-agent interoperability, federated collaborationNot a replacement for internal orchestration or direct tool invocation

MCP Strategy

What MCP is

An open protocol for connecting AI applications to external systems through a structured client-server model based on JSON-RPC.

Why MCP matters

Provides a consistent, discoverable way for LLM-backed agents, copilots, and external AI clients to access internal capabilities without one-off integrations.

How Synedrix OS uses MCP

MCP is adopted at the tool abstraction boundary, not at the orchestration core. Recommended patterns:

  • Expose selected internal services as MCP servers.
  • Publish safe read-only resources first.
  • Introduce MCP tools only when approval and permission semantics are clear.
  • Route MCP-exposed tool execution through the same governance and observability controls as native tool calls.
  • Trace MCP-backed LLM activity through Langfuse and platform observability.

First candidates for MCP exposure

Knowledge base and architecture documentation resources; schema discovery for canonical data models; read-only analytics and reporting tools; operator assistant queries; selected domain-pack retrieval resources. Write-capable tools come later.

A2A Strategy

What A2A is

An open communication protocol for interoperability between AI agents and agent systems, enabling secure cross-system task exchange.

Why A2A matters

Synedrix OS is a reusable operating substrate, not a sealed ecosystem. A2A enables cross-system agent delegation, external specialist-agent collaboration, federated workflows, and standardized communication with third-party agent ecosystems.

How Synedrix OS uses A2A

A2A is adopted at the platform edge, not as the internal coordination mechanism. Recommended patterns:

  • Publish selected Synedrix OS agents or workflows as A2A-compatible endpoints.
  • Consume external A2A agents as bounded collaborators, not trusted internal peers by default.
  • Treat every external A2A interaction as policy-governed and auditable.
  • Preserve correlation IDs and workflow lineage across A2A boundaries.
  • A2A-originated tasks must still pass through Synedrix OS governance and approval controls before side effects.

MCP and A2A Together

Complementary, not competing. Use MCPwhen an agent needs structured access to a tool, prompt, or context resource. UseA2A when one agent system needs to delegate to, collaborate with, or request results from another agent system.

Relationship to Existing Layers

  • Orchestration: n8n remains the internal orchestration kernel (Layer 3); external interoperability protocols live at the platform boundary.
  • Agent Runtime: LangGraph is the Agent Runtime (Layer 4); n8n invokes it as a governed internal step (see the boundary section below).
  • Tool Abstraction: MCP is one standard contract model inside this layer; native adapters remain valid.
  • Governance: All protocol-mediated interactions must be policy-aware.
  • Observability: MCP tool calls, A2A task exchanges, and protocol-level failures must be traceable.

The n8n ↔ LangGraph runtime boundary

Per ADR-OS-001, complex agent reasoning runs in the LangGraph Agent Runtime (Layer 4), invoked by n8n (Layer 3) as a governed step. This is an internal boundary — it is not cross-system A2A federation — but it deliberately reuses the same transport and governance primitives:

  • Transport: HTTP for internal service calls; MCP when the runtime is consumed as a tool/context server; A2A only when the runtime is delegated to or from a separate agent system at the platform edge.
  • Lineage: the invocation carries correlation_id, causation_id, the resolved active_tool_set, and deployment_mode; LangGraph returns the canonical Outcome Envelope and emits os.outcome.* events. No lineage or observability gap is permitted at the boundary.
  • Governance: Policy Guard (os-agt-002) and HITL Gateway (os-agt-004) remain reachable from inside LangGraph; the same “no protocol bypass of governance” principle applies. The boundary does not introduce a new trust tier — internal invocation is governed as a native tool call, not as an untrusted external A2A partner.

The full boundary topology and governance bridge are specified in the Orchestration & Runtime Strategy §2.1 and §5.

Security and Governance Principles

  1. No protocol bypass of governance.
  2. Least privilege by default.
  3. Read-first exposure model.
  4. Observable by default.
  5. Versioned contracts.

These principles are operationalized per agent in the mcp_tool_policy and a2a_tool_policy blocks of the Agent Runtime Policy Spec. See Synedrix OS Agent Runtime Policy Spec v0.3 — Tool Authorization for the full three-layer model (capability ceiling → runtime restriction → instance override) that implements “least privilege by default” and “no protocol bypass of governance” at the per-agent level.

Adoption Roadmap

  1. Phase 1 - Standards-aware internal architecture: define stance, identify MCP exposure candidates, define external-agent trust model.
  2. Phase 2 - Controlled MCP enablement: publish a small number of internal resources, add tracing/permissions/audit hooks.
  3. Phase 3 - Bounded A2A interoperability: expose selected workflows, integrate one external agent system in a low-risk flow.
  4. Phase 4 - Federated protocol ecosystem: expand across domain packs, introduce protocol catalogs, trust tiers, governance templates.

Non-Goals

This strategy does not imply replacing internal orchestration with A2A, replacing native adapters with MCP everywhere, exposing every internal service publicly, or assuming protocol support automatically solves governance, trust, or safety.

Positioning

Synedrix OS is an open, standards-aware agentic operating system. MCP is the preferred protocol for agent-to-tool/context integration; A2A is the preferred protocol for cross-system agent interoperability. Both are governed interoperability layers that extend the OS without replacing its orchestration, governance, or observability core.

Versioning

  • Version: 0.3
  • Supersedes: v0.2

Canonical Reference

Orchestration and Runtime

Architecture

Deployment

Agent Specifications

  • Synedrix OS - Agent Spec — Full agent anatomy; the A2A communication model in §A2A implements this strategy at the agent level.

Synedrix OS · April 2026