Systems & Control

Systems & Control

Precision agriculture is an act of faith—trusting data, sensors, and algorithms to act as extensions of human intent. FarmCore and FarmLytics were designed for efficiency, not exploitation. What follows illustrates how their interaction should have worked—until trust itself became the point of attack.

These diagrams are reconstructions, not warnings. They show the hinge where design met assumption.

Attack Summary (System View)

  1. 1 Trust Boundary — Cloud → Site → Edge Boundary
    Certificates and policies anchor trust across FarmLytics, Gateway, and FarmCore.
  2. 2 Policy / Model Injection Injection
    Poisoned weights and spoofed certs traverse the update and trust channels.
  3. 3 Recipe Override at Gateway Control
    Gateway relays “optimized” recipes; FarmCore shadow test passes under spoofing.
  4. 4 Telemetry Spoofing HMI
    Operators see healthy signals while actuators drive plant stress.
  5. 5 Safety Kernel v2 Mitigation
    Dual‑channel physics checks and local policy verification harden the edge.

HMI / CEA Recipe Dashboard

Reconstructed interface view. Simulated data for educational context.

Surveillance feed: Vertical Farm Sector A
CAM 03 — Vertical Farm Sector A
Surveillance feed: Main Floor North - Low Light Mode
CAM 01 — Main Floor North (Low Light)

Launch the dashboard →

Case Study — Stuxnet

In 2010, a piece of code slipped through the world’s air gaps and rewrote how we think about control. Stuxnet targeted Siemens industrial controllers at Iran’s Natanz facility, quietly varying centrifuge speeds while feeding normal data back to operators. It was the first malware to cause physical destruction—proof that software could move steel.

To engineers, it was a revelation of possibility and responsibility. The systems behind Dark Recipe draw their lineage from that moment: a world where process automation and statecraft share the same syntax.

“We built it to keep the peace. We just forgot to teach it what peace meant.”