Skip to content
Technology How it works Breeder — Hyperion Burner — Aegis Burner — MetroVolt AI-Native Architecture Magnets Fuel cycle Safety Roadmap
Solutions AI & Data Centers Defense & Government Grid & Baseload Neutron Detection Quantum
Learn Technical Library
Proof Publications Whitepapers Technical Library Open Science & Reproducibility The Honest Gates
Company About / Mission Leadership Environment Health & Safety Investors Careers Press Contact
3D Model
AI Architecture › L5 · Applications & Copilots
L5 · Applications & Copilots

The Operations Copilot: Scope and Reasoning Loop

An advisory agent that generates runbooks and procedures, supports shift handoff, and assembles campaign context from the twin and the data fabric.

THE STACK · click to jumpL7Ecosystem & StrategyL6Experience & VisualizationL5Applications & CopilotsL4OrchestrationL3Twin Modeling & AIL2Data FabricL1Control PlaneL0Foundation▲tlmctl▼L5 · APPLICATIONS & COPILOTSAgentic copilots that reason over the machine.1Plasma Copilotscenario design2Engineering Copilotsubsystem analysis3Operations Copilotrunbooks & procedures4Agentic Toolsbounded action-taking5Knowledge BaseRAG over the fabric6Guardrailssafety-boundedMACHINE TIEReads the twin and fabric; proposes actions that route through L4.KRONOS FUSION ENERGYAI-NATIVE S.M.A.R.T. GENERATORAPPLICATIONS & COPILOTSSHEET 07REV. 2026-08L5 · AI-NATIVE STACK
L5 · Applications & Copilots — its place in the stack (left, click any layer) and its internal components (right). Telemetry rises; control descends.

What the Operations Copilot does

The Operations Copilot is the procedure-facing agent of L5. It converts approved scenarios and maintenance plans into executable runbooks, generates step-by-step operating and recovery procedures, supports shift handoff, and assembles campaign context. It is the connective tissue between the Plasma and Engineering copilots' proposals and the humans who carry them out.

Three capabilities

Every generated artifact is grounded in retrieval over approved procedure templates, standing orders, and the current twin state, and every step that has a plant side-effect is marked as requiring L4 authorization and human confirmation. The Operations Copilot writes procedures; it does not run them.

Grounded, not invented

A generated runbook is assembled from a retrieved template plus current context, never freely composed. Each step cites its source template and any twin-state precondition, so a reviewer can confirm the procedure matches approved practice. Steps that would exceed the safe operating envelope are refused at generation time.

text
generate_runbook(approved_scenario):
  tmpl  = rag.retrieve('runbook template', approved_scenario.type)
  state = twin.current_state()
  steps = fill(tmpl, approved_scenario, state)
  for s in steps:
     s.precondition = check(state, envelope)
     s.authorization = 'L4 + human' if s.side_effect else 'none'
  return runbook(steps, cite=[tmpl, state])

The following pages detail runbook generation, procedure generation, shift handoff, and commissioning support. Because both machines are pre-FOAK design-and-simulation studies, the Operations Copilot's near-term product is the commissioning and first-campaign procedure set validated against the twin.

Content reviewed August 2026 · design-and-simulation stage