Back to catalog
LabpremiumDistributed Intelligence Units

Lab: Build a Simple Agent That Runs in the Mesh

Build and deploy a DIU that functions across a distributed mesh, managing state and execution through policy-driven routing.

advanced 120 minutesFree

What you'll build

Vertical stack context: DOSFIUniversity (research) → DOSFI Academy (this course) → MeshNativeExchange (publishing & monetizing). Students progress from Learning → Building → Publishing → Monetizing. In this lab, you will architect a Distributed Intelligence Unit (DIU). You will progress through the lifecycle of a mesh-native agent: defining its identity via a strict manifest schema, implementing a perception-reasoning-action loop, and validating its execution path across disparate network nodes. This lab emphasizes architectural rigor and sovereign system control within a closed-core environment.

Skills you'll practice

  • Define a DIU manifest with identity, capability, and power budget constraints.
  • Implement a perception-reasoning-action loop logic.
  • Establish secure connections to ingestion and execution nodes within the mesh.
  • Implement a deterministic routing policy for internal agent telemetry and inference.

Before you start

  • Understanding of mesh network topologies.
  • Proficiency in JSON schema design.
  • Familiarity with distributed state management principles.

Lab steps5 steps

  1. 1
    Defining the DIU Manifest
  2. 2
    Implementing the Perception-Reasoning-Action Loop
  3. 3
    Lesson 3: Connecting the Distributed Intelligence Unit to Mesh Nodes
  4. 4
    Lesson 4: Implementing Intelligent Routing Policies
  5. 5
    Lesson 5: Simulating Agent Cycles and Tracing Execution