Skip to content
Team Acalvio
|
September 23, 2026

OT cybersecurity risk management: an agentless deception playbook

An intruder who reaches a plant’s control network does not need a zero-day. In January 2024, FrostyGoop malware sent ordinary Modbus TCP commands to heating controllers in Lviv and cut heat to more than 600 apartment buildings. OT risk management cybersecurity programs exist to price that kind of consequence before it happens. Managing OT cyber risk means ranking assets by the safety, availability, and business impact of their compromise, documenting credible attack scenarios, and treating the highest risks with controls that operators approve and that do not disturb the process. Agentless deception is one detection layer that fits inside those limits.

At a glance

  • Definition: OT cyber risk combines asset criticality, threat exposure, likelihood, and safety, availability, and business consequence.
  • Workflow: Inventory, scenario, evaluation, treatment, owner, and review date, with testing only through approved, non-disruptive procedures.
  • Worksheet: A four-column risk register links each scenario to evidence, a control, and an accountable decision.
  • Deception’s role: Isolated decoys give high-fidelity leads without agents on controllers. In the U.S. Navy ANTX FY25 exercise, Acalvio deception produced 100% true positive alerts and denied 80% of attacker objectives against automated, credential-driven techniques in the exercise.
  • Limit: Decoy activity informs risk decisions; it is not proof of risk reduction.

What is OT cybersecurity risk management?

OT cybersecurity risk management is the continuing process of identifying, ranking, treating, and reviewing the cyber risks that could affect safety, availability, and production in operational technology (OT) and industrial control system (ICS) environments.

OT risk has six inputs. Asset criticality describes how much a process depends on a controller, workstation, or network path. Threat exposure covers how an intruder could reach it, through remote access, the IT/OT boundary, removable media, or a supplier connection. Likelihood weighs that exposure against existing controls. Safety impact, availability impact, and business consequence describe what happens if the asset is misused, and in OT the first of those usually carries the most weight.

An OT cybersecurity risk assessment is a point-in-time evaluation of those inputs. Risk management is what follows: assigning treatments and owners, tracking evidence, and revisiting decisions when the plant, the threat, or the controls change. The decision-makers include the plant or asset owner, process safety, OT engineering, and security, and none of them can make the call alone.

OT risk management vs. IT security assessment

OT and IT risk work share a vocabulary but rank different things first. OT puts safety and uptime ahead of confidentiality, which changes what a team can safely test, deploy, and do during a response.

FactorIT security assessmentOT risk management
Top priorityConfidentiality, integrity, and availability of dataSafety, then availability of the physical process
EquipmentRefreshed on short cycles; agents and patches routineLegacy controllers in service for decades; agents often unsupported
DiscoveryActive scanning is common and generally safePassive-first discovery, because active probes can disrupt fragile devices
Change windowsFrequent and flexibleTied to maintenance outages and plant change control
OwnershipIT and security teamsPlant owners, engineering, and process safety, with security advising
ResponseIsolate or reimage hostsAny action that touches the process needs operator approval

The practical consequence is simple. An IT scanner or endpoint agent that works well on corporate laptops may be unsafe on a controller, so OT teams depend on passive and out-of-band controls.

Build an OT cyber-risk management workflow

A workable OT cyber-risk program runs in seven steps, each with an owner and a point at which it stops or escalates. This is an editorial framework that organizes practice; it is not a published standard.

StepInputOwnerOutputReview or stop condition
1. Inventory approved assetsPassive discovery, engineering drawings, existing recordsOT engineeringAsset list with named ownersStop if discovery would require active probing that the owner has not approved
2. Identify critical processesProcess descriptions, safety studies, dependenciesPlant owner and process safetyRanked critical processes and supporting assetsReview when the process changes
3. Document credible scenariosThreat intelligence, remote-access paths, past incidentsSecurity with OT engineeringWritten attack scenarios per critical processDrop scenarios with no plausible path
4. Evaluate likelihood and consequenceScenarios, existing controls, safety impactRisk ownerRated risks in the registerEscalate any risk with a safety consequence
5. Prioritize controlsRated risks, control options, outage windowsRisk owner and engineeringTreatment planDefer only with a documented acceptance
6. Assign an owner and review dateTreatment planRisk ownerAccountable owner and date per riskOverdue reviews go to leadership
7. Test controls safelyApproved test procedure, rollback planOT engineering and securityEvidence that a control works as designedStop on any unexpected process behavior

Step seven matters most in a plant. A control that cannot be tested without touching production is tested in a lab, on a spare, or during an outage.

Where agentless deception can support OT risk reduction

Agentless deception supports OT risk reduction by producing a high-fidelity signal from assets that no operator, engineer, or control process should ever touch, without installing software on controllers or placing an appliance in line with production traffic.

Passive OT monitoring tools see a great deal, but much industrial traffic is legitimate by design. An engineering workstation writing to a controller looks the same whether a technician or an intruder is at the keyboard, so intent is hard to judge in isolation. Deception changes the question to a simpler one: why did anything touch an asset no legitimate process should need?

Acalvio’s published OT guidance places decoys near remote-access termination points and at the IT/OT boundary, which the Purdue model calls Level 3.5, and pairs low-interaction decoys for scale with higher-interaction decoys built from custom web interfaces and golden images. Deceptive credentials and breadcrumbs on the IT or OT side point an intruder toward those decoys and carry less production risk than agents. Hand-built honeypots cannot keep up with automated, credential-driven intrusions, so automation is the entry requirement: the ShadowPlex Preemptive Cybersecurity Platform deploys and adapts deceptive assets autonomously, and 360 Deception, Acalvio’s three-vector framework, aims to detect, divert, and degrade the attacker. For more detail, see Acalvio ShadowPlex in OT/ICS environments.

Deception runs alongside passive OT monitoring, segmentation, SIEM, EDR on supported hosts, IAM, PAM, and MFA rather than replacing them. A decoy interaction is a lead for investigation; authorized scanners and vendor tools can reach decoys too.

The SANS Institute’s Five ICS Cybersecurity Critical Controls (Lee and Conway, 2022) offer a practical way to place deception in a program. The mapping below is illustrative, not a certification or compliance claim.

SANS critical controlWhere deception can contributeWhat other controls still own
ICS incident response planHigh-fidelity alerts that start a defined, safety-aware triage pathThe plan, roles, and process-safety decisions
Defensible architectureDecoys at Level 3.5 and inside zones show whether segmentation holds against real attemptsSegmentation, firewalls, and industrial DMZ design
ICS network visibility and monitoringA behavior-independent signal next to passive monitoringAsset inventory, protocol analysis, and baselining
Secure remote accessDecoys and deceptive credentials near remote-access termination pointsMFA, jump hosts, session brokering, and access approval
Risk-based vulnerability managementEvidence of which exposures intruders actually probeVulnerability inventory and patch or mitigate decisions

OT risk register and control-mapping worksheet

An OT risk register earns its place when every row connects a scenario to evidence, a control, and an accountable decision. The rows below are illustrative planning prompts, not verified deployments or customer results. Validate each entry with plant engineering and the control owner.

Risk scenarioEvidence to collectControl to assessOwner and decision
Unexpected access to a critical zoneAsset ownership, access paths, approved communicationsSegmentation and approved monitoringOT owner confirms scope and a safe investigation path
Misused maintenance accessIdentity records, maintenance window, remote-session contextAccess governance and monitored deceptive resources where safeJoint IT/OT review before any response
Unmonitored legacy equipmentPassive inventory, criticality, data-coverage gapsNon-disruptive visibility and validated agentless optionsPlant engineering approves the design
Slow escalation of suspicious activityAlert history, contact tree, existing response procedureDocumented triage and safety-aware responseRisk owner assigns a review date

Each decision establishes who acts and under what limits. It does not establish that a risk is reduced; that takes later evidence. Return a row to the owner whenever the evidence is incomplete or the treatment would touch the process.

Illustrative OT risk-review scenario

This scenario is hypothetical. It shows how the register works, not a real customer event.

Objective: Test the “misused maintenance access” row for a packaging line whose vendor connects through a jump host in the industrial DMZ.

Observation: At 2:00 a.m., outside any scheduled maintenance window, a deceptive credential placed on the jump host is used to log in to a decoy engineering workstation in Level 3. The decoy is isolated, so the security team can investigate it freely.

Corroborating evidence: The team checks the maintenance ticket system, the vendor’s remote-session logs, and identity records. The session belongs to the vendor account, but no ticket covers it.

Benign alternative: A vendor technician may have run a scripted discovery tool during a window that was approved verbally but never recorded.

Safe triage: Nobody queries or reconfigures a production controller. Operators confirm the line is running normally, and the OT owner decides whether to suspend the vendor account.

Decision: The vendor confirms no technician was working. The account is suspended pending review, the row’s likelihood rating rises, and the treatment becomes approved session brokering for all vendor access, with a review date in 30 days.

Implementation and operational guardrails

Introducing deception into an industrial network is a change to that network and goes through the same approvals as any other change. Do not run active tests on safety-critical systems.

Before deployment, confirm this readiness checklist:

  • Plant-owner approval: the asset owner signs off on decoy placement and purpose.
  • Protocol validation: Acalvio and plant engineering confirm which industrial protocols the decoys emulate in this deployment.
  • Isolation: decoys cannot send commands to, or receive traffic intended for, production devices.
  • Passive-first discovery: inventory comes from passive sources before anything is placed.
  • Change-control review: the deployment follows the plant’s change process, with a rollback plan.
  • Response boundary: a written line between IT investigation and process-control actions, with operations owning the latter.

Assign each item an owner, a validation step, an escalation path, and a review date. Product-specific implementation steps need Acalvio technical review.

Measure risk-treatment progress

Measure whether risks are owned and treated, not how many alerts fire. Set a baseline for each measure, name its data owner, and define its calculation before reporting it.

MeasureWhat it showsWhat it cannot show
Critical assets with a named ownerAccountability coverageWhether owners act
Risks with an assigned treatmentRegister disciplineWhether treatments work
Evidence gaps closedBetter data for decisionsFewer intrusions
Approved control reviews completedControls tested as designedCoverage of untested paths
Escalation-readiness checks passedResponse paths work in drillsPerformance in a real incident

Decoy interaction counts belong in investigations, not in progress reports. A quiet decoy may mean good segmentation or an intruder who has not arrived yet.

The path forward

OT risk management cybersecurity work comes down to a few disciplines: rank assets by safety and availability consequence, write credible scenarios, assign owners and review dates, and test controls only in ways the plant approves. Agentless deception adds a high-fidelity, behavior-independent signal that fits those limits without agents on controllers. The goal is to stop asking analysts to infer intent from weak evidence when the environment can produce a stronger signal.

See how Acalvio approaches OT and ICS security with agentless deception, or read how FrostyGoop abused Modbus TCP and where OT decoys fit. For the federal guidance most OT programs build on, see NIST SP 800-82 Rev. 3, Guide to Operational Technology Security.

Frequently asked questions about OT cyber risk

Inventory approved assets and their owners, identify critical processes, and document credible attack scenarios. Rate each scenario by likelihood and by safety, availability, and business consequence, then assign treatments, owners, and review dates. Test controls only through approved, non-disruptive procedures.

No single method fits every plant. A strong ICS risk assessment starts from passive asset discovery and safety studies, uses scenarios tied to real access paths, and involves plant engineering and process safety in every rating. Frameworks such as NIST SP 800-82 and the SANS Five ICS Critical Controls help structure the work.

Decoys and deceptive credentials sit beside production systems rather than on them, so nothing is installed on controllers. Because no legitimate process needs those assets, an interaction is a high-fidelity lead that feeds investigation and the risk register. Deception adds detection; it does not replace segmentation, access control, or monitoring.

OT ranks safety and process availability first, runs legacy equipment that often cannot take agents or scans, and changes only within maintenance windows. Response actions that touch the process require operator approval.

Plant-owner approval, protocol validation, isolation from production devices, passive-first discovery, change-control review with a rollback plan, and a written boundary between IT investigation and process-control actions. Product-specific deployment steps need vendor technical review.

Content
Acalvio, the Ultimate Preemptive Cybersecurity Solution.