Case Study

Governing a physical farm-control prototype before execution.

Sentinel SCA was tested around an ESP32-based farm-control prototype connected to Agri-Nexus Prime prototype work. The goal was to prove that autonomous commands can be checked before they reach real actuators.

View Demo ยท

Guided prototype video showing sensor readings, ESP32 wiring, and fan control.

Watch Prototype Video

Problem

Autonomous farm-control systems can trigger physical actions such as cooling, relay switching, valve logic, or feed automation. Unsafe actions need a control boundary.

Prototype

The testbed used ESP32 hardware, sensors, relay/fan control concepts, and Sentinel oversight before controlled execution.

Sentinel Role

Sentinel validates signed commands, returns allow/deny/review decisions, and records audit evidence.

Governance Flow

  1. Sensor or agent proposes an action.
  2. Command is signed by an agent identity.
  3. Sentinel checks identity, replay, policy, and risk.
  4. Only approved actions proceed toward controlled execution.
  5. Decision evidence is stored for tenant audit and export.

What It Proves

  • Sentinel can govern more than dashboards or logs.
  • Physical-world automation can be placed behind a policy boundary.
  • Operators can see evidence after every decision.
  • The same pattern applies to AI agents, infrastructure automation, and IoT control.

Evidence Produced

The prototype demonstrates the same evidence model customers use in Sentinel: signed agent identity, decision result, risk score, reason, timestamp, tenant audit timeline, and exportable evidence package.

Ed25519 Identity Replay Protection Tenant Audit Evidence ZIP Ops Visibility