Skip to content

Access control for the machines

Know who enters your factory and control what machines they may power and use. Authorise it, block it, and prove it, even without any available network.

Most industrial equipment has no login, no screen and no IP address. It has a key switch, a padlock, and a paper logbook. Trust-Buster adds a trust layer around it without opening it, modifying it, or connecting it to any network.

Trust-Buster is developed in France, in Sophia Antipolis (06), on a very specific segment of OT cybersecurity: hardware and physical protection as close as possible to the equipment.

Where traditional players secure network flows or operating systems, Trust-Buster operates on the raw hardware (power sockets, PLCs, off-network machines, USB/serial interfaces) that is unable to host a conventional software agent.

Key features and building blocks

  • Physical and local access control (“the Cube”): a standalone central unit that validates the operator’s actual presence, identity and authorisations right at the heart of the machine room, without depending on a cloud connection or a remote server.

  • Hardware interception (the Satellites): physical units inserted on USB ports, serial links or power sockets. Controlled by the Cube, they block hardware attacks such as keystroke injection (BadUSB) or unauthorised tampering with interfaces.

  • Management in an isolated environment (Air-Gap): updating schedules and policies, and retrieving encrypted audit logs, are carried out offline via a secure mobile terminal.

  • Radio anomaly detection: local identification of unauthorised devices nearby (AirTag, suspicious signals, radio jamming) in order to raise an alert or trigger protective measures.

Position in your architecture: it is a complementary solution that sits upstream of levels 0 and 1 of the Purdue model (actuators, sensors, PLCs), particularly suited if you need to secure physical access by external contractors on critical or isolated sites.

Trust-Buster for Compliance and Cybersecurity

  • Why you need it

    Friday, 23:20. A five-axis machine starts up in a shared workshop. It runs for forty minutes, produces eleven parts, and stops. On Monday, someone notices.

    Who started it? The building badge proves someone entered the building. It says nothing about the machine. The machine's hour counter says forty minutes — unattributed, unsigned, reset at the last service.

    With Trust-Buster, starting that machine requires three things at once: the person is physically present, their identity is proven by hardware, and the intervention was planned. All three, or nothing. The log entry is signed: who, which machine, when, under which rule, allowed or denied.

  • What you actually buy

    The hardware on your site

    The Cube — the decision unit. A fixed appliance with radio sensors and communications, installed per room, per building or per cabinet. It holds the policy, evaluates it locally, signs the log, and commands the effectors. It carries a secure element for hardware-rooted identity.

    Effectors — the enforcement points. Small units that sit outside the machine, on its power cable. One per controlled outlet. They enforce what the Cube decides and nothing else: the physical button is decoupled from the relay, so pressing it signals intent rather than issuing a command.

    Credentials. A radio badge carried by the person, and a mobile app for approval workflows and for carrying policy updates and logs to and from offline sites.

  • How it decides

    The Cube establishes three certainties before authorising anything. Any one of them missing means denial.

    Detection. The Cube continuously observes its radio environment and knows which enrolled devices and people are nearby. This is an entry condition, not an identity claim.

    Identity. A cryptographic challenge answered by a key held in a secure element. Not a number broadcast by a badge, and not something an imitation produces.

    Validation. The right person is still not enough. The Cube checks that this action, on this machine, at this time, matches a planned intervention.

    Actions. The effector releases or cuts power, verifies the resulting state, and logs it. Unplugging, rewiring or resetting a controlled outlet is itself an event: the zone locks, the incident is recorded with its reason.

Navigating Regulatory Challenges with Ops-Devices Systems

Understanding the Regulatory Context

NIS2 and the Cyber Resilience Act (CRA) make asset inventory, OT risk management and incident evidence mandatory, with heavy penalties. Trust-Buster, the system built on the Ops-Devices platform, turns these obligations into an operated, on-site capability: inventory, access control and a signed audit trail, managed for you. You stay focused on operations and your compliance evidence writes itself.

Ops-Devices: trust, anchored in hardware

Our platform empowers Trust-Buster operations, relying on a physical appliance, the Cube, as a communications & sensors gateway that secures any equipment even if it has never been designed for authentication. It recognizes what and who is actually present, only authorizes actions from an approved set of rules, and acts on its own — on site, even when disconnected from any network. Any attempt at physical bypass is detected, the perimeter locks down, and the incident is logged.

Designed and developed in Europe. Operational sovereignty: no data leaves your site, no third-party cloud, no remote kill-switch, full air-gap operation. Our supply chain is publicly documented.

Who Trust-Buster is for

For industrial operators and critical infrastructure. Factories, energy, water, transport, sites subject to NIS2 and IEC 62443. You operate equipment that was not designed with cybersecurity in mind and cannot simply be replaced. Your supervision teams (SOC) need to extend their visibility and control all the way to the physical boundary — where software tools stop.

For field operators and defense environments. Sites without reliable connectivity, by constraint or by doctrine. You need a trust chain and an action capability that work autonomously, without relying on the cloud, and whose sovereignty is non-negotiable.

More broadly, for anyone who must control “who can activate or use which equipment, where, and when” — and know, without ambiguity, when someone attempts to bypass it.

Our key assets

  • Design strengths

    Five decisions, and what each one costs

    Every architecture is a trade-off. Here are ours, with the bill attached.

    Identity lives in hardware. A badge answers a cryptographic challenge from a secure element certified Common Criteria EAL5+ — a certification carried by the component itself. No shared secret on a server to steal. Cost: a lost badge is revoked, not reset.

    The decision is taken on site. No cloud, no server that has to answer for a machine to start. A site cut off for three days runs exactly as designed. Cost: no live dashboard — policy and logs travel by phone.

    Enforcement sits outside the machine. Effectors clip onto the power cable. Nothing installed inside, CE marking intact, removable without a trace. Cost: we govern the path to the machine, not what happens inside it.

    It fails to denial. Every failure lands de-energised; no breakdown opens an access. Cost: someone who wants to stop your line rather than get into it can jam the band. Better said now than discovered in month three.

    Built in Europe, sourcing published — including the decisions we reversed when a supplier changed hands. Cost: a narrower supplier list, and sometimes a slower roadmap.

    None of the five is unique alone. We have not found another system that makes all five at once — if you know of one, we would like to see it.

    Physical access remains physical access, and no radio monitoring proves the absence of a signal it never saw. What changes is that an attempt stops being invisible.

  • Proof demo on request

    See if you can break it

    We will not describe our security to you. We will let you try to break it.

    You take the intruder's seat, facing a real machine on a controlled outlet. No script, no briefing, no demo mode. Press the outlet's button, unplug it, plug the machine in elsewhere, rewire it directly, reset the unit, show up without a valid identity, ask for access outside your window.

    At each attempt the system reacts in front of you — cut-off, alert, lockdown — then tells you what it saw in your gesture and why it refused. Some attempts will work. We show you those too, and what we are doing about them.

    You leave with a signed audit log of your own attempts: your future incident report, written live by a volunteer attacker.

    Practical: around 90 minutes, at your site or ours. Bring your most sceptical colleague, and your own tools if you have them. Useful attendees: whoever owns the machines, whoever owns compliance, and whoever will live with the false positives. Nothing is installed on your equipment.

Trust-Buster : See it to believe it

Playing loads the YouTube player.

Team

Pascal FLAMAND

Pascal FLAMAND

Lead

CEO and founder @JANUA - Now Open-IAM - since 2004, a company specializing in IAM (security, access control, identity management) and Open Source.

CEO and Founder @Samarcande since 2023, holding, acting as an investor in real business projects.

Julien HOLTZER

Julien HOLTZER

Technical Advisor

Passionate about electronics, embedded devices, research and development, building new hardware and software disrupting solutions.

15+ years of experience in Smart Cities and Smart Buildings, from concepts to implementations.

Frequently Asked Questions

Why Trust-Buster? Busting trust?

Until now, Zero Trust stopped at the network. We extend it all the way to the power socket.

Zero Trust – “never trust, always verify” – is the reference model that every cybersecurity decision-maker seeks to deploy. Yet until now it stopped at the logical layer: faced with an electrical cabinet, an industrial PLC from 2009 or a simple wall socket, this paradigm simply no longer existed.

Our approach is to bridge this gap: we never rely on the trust declared on a package or in a datasheet. Our devices are designed to continuously monitor users and hardware alike. They provide proof that operations are running correctly and detect any action likely to break this chain of trust – whether it is a handling error or a malicious act.

Is Trust-Buster right for my needs?

We designed Trust-Buster to address demanding critical situations: high-risk equipment, sensitive perimeters, stringent compliance obligations and traceability that stands up to any auditor. But you can also deploy it to better understand and analyse your equipment fleet: what power consumption, which permanent equipment is under-used, which temporary uses fall outside your limits.

Trust-Buster is deployed in three modes, and you decide when to move on to the next one, equipment by equipment, never globally.

Observation: it detects, alerts and logs without ever cutting power, notably during the first weeks of deployment. You measure the real noise before arming anything.

Consent: it asks a manager to approve or refuse an access request from their phone. Without a response, nothing moves: ideal for shared equipment, during contractor interventions, or for high-turnover sites without established rules.

Enforcement: it cuts power according to the rule, without human intervention. This requires an explicitly listed perimeter, on equipment whose shutdown has no consequences.

Most of our deployments are not meant to move to Enforcement everywhere: a site under Observation already produces the audit log that lets you better understand and react on a case-by-case basis.

What scenarios did you design your product for?

The origin of the Cube goes back to the first time we took on criminal liability for the use of machine tools in a FabLab, in 2015. You no doubt face the same doubts, and they cannot be settled with a “Trust me”. Here are the four questions that cost dearly:

“Who started the press at 3 a.m.?”

The situation: a restricted-use machine on a shared site, an incident occurs, with no witness.

Today: the access badge proves an entry into the building, and is very easily bypassed. It does not tell you who started the machine.

With Trust-Buster, start-up requires presence, hardware-proven identity, and a scheduled intervention. All three, or nothing. The log entry is signed: who, which equipment, when, under which rule, authorised or refused. The departure or absence of an authorised user is also detected, and an alarm indicates that the machine can still be activated, or the machine is even electrically protected against being restarted.

“Was this person authorised?”

The situation: co-activity, with cascading interventions by outside contractors, and an authorisation that expired three weeks ago.

Today: the answer is in a binder, reconstructed after the fact, and it will not stand up in front of an auditor.

With Trust-Buster: the authorisation is a rule, not a document. Once expired, it no longer authorises anything. A refusal is traced just like an approval. You obtain an enforceable list of granted and refused accesses, per contractor and per piece of equipment.

“The equipment was rewired live over the weekend.”

The situation: a power socket is moved, a power strip is added, or a control unit is reset.

Today: software supervision sees nothing. The tampering is discovered at the next incident.

With Trust-Buster: every unplugging, rewiring, reset and abnormal consumption is a security event. The perimeter locks down, the incident is logged. The timestamp of the bypass is directly accessible, and you can define a response rule requiring that you be contacted before the equipment is put back into service, even after the original wiring has been repaired.

“The site has been cut off from the world for three days. Is security still holding?”

The situation: your site is isolated, with no conventional IT link any more.

Today: IT security is adapted to work with OT equipment but depends heavily on the network, or even ends up relying on the cloud. In practice, it gets disarmed and work carries on, because production comes before IT.

With Trust-Buster, this is normal operation. The Cube decides on its own. Rules come in and logs go out via the phones of authorised and authenticated carriers.

The complete log for the isolation period is available, signed, with no gaps.

And when things go wrong?

We do not cut off a machine mid-cycle

By default, enforcement applies to start-up authorisation, not to interruption. Equipment that is already running triggers an alert, not a shutdown, unless it explicitly appears on the list of equipment that may be cut off while running, which you draw up and sign with us.

Degraded mode is a mode, not an accident

  • Cube out of service: the satellites keep their last authorised state until their capacity runs out. No cascading shutdowns.

  • Satellite out of service or unplugged: the event is logged and reported. The equipment is not powered back on without the system’s knowledge.

  • No network: this is normal operation, not a failure. Decisions are local and offline by design.

  • Mains power cut: when power returns, the enforcement points restart de-energised. Putting equipment back into service is an authorised act, never a side effect.

Break-glass: a traced way out

A designated manager can force access in an emergency through an offline procedure: a paper QR code sealed in an envelope placed in each room, scanned and validated on their phone. Access is then granted immediately, and is at once timestamped, signed and reported in the first report.

We prefer a traced exemption to an invisible bypass. It is the same principle as the rest of the system.

We count false positives

A security device that regularly blocks wrongly, even without any significant incident, gets unplugged within six months out of sheer exasperation. That is what Observation mode is for: before anything is armed, it produces the real rate of events per day and per piece of equipment, on your site. If the figure is not sustainable, you do not switch that equipment to Enforcement. You then refine the rules, narrow the perimeter, or give up systematic control of that equipment (it remains in the log). That is a decision, not a failure.

Who is responsible?

The rule is yours: you define it, you validate it, it is signed. We guarantee its faithful execution and the associated proof. The decision to arm a piece of equipment in Enforcement mode remains with the operator, as with any lockout device. The service contract states this in black and white, and names the safety and security contact on your side

What Trust-Buster brings to your compliance

No product makes an organisation compliant. A product provides evidence, and covers certain measures.

Under NIS2: your obligations as an operator

  1. Asset inventory and control. The Cube keeps the inventory of protected equipment and enrolled carriers, and flags the appearance of any undeclared transmitter within the perimeter. It is a field inventory, including for equipment that your asset management tool cannot see because it is not on the network.


  2. Access control, including for contractors. An authorisation becomes a dated rule, not a document in a binder. A granted access and a refused access are traced on the same footing, and you produce an enforceable list of interventions, per contractor and per piece of equipment.

  3. Traceability and incident management. Every authorisation, every refusal, every bypass attempt produces a timestamped, justified and signed entry, exportable to your SIEM without necessarily being connected to it at all times. It is the raw material for an incident notification, not the notification itself.

  4. Continuity in the event of unavailability. Decisions are made locally. A network outage, a failure of your core IT or an attack targeting your infrastructure changes nothing in how the device behaves in the field.

Under IEC 62443: the technical framework, voluntary but expected

IEC 62443-4-2 explicitly provides for physical access control as a compensating countermeasure when industrial equipment is unable to identify a person individually.

Our architecture follows the design principles of IEC 62443-4-2, with a target level of SL-C 2. No third-party assessment has been carried out to date. The Common Criteria EAL5+ certification we mention applies to the secure component we integrate, not to our product.

Under the CRA: our obligation as a manufacturer

The CRA imposes nothing on you as an operator. It requires us, as the manufacturer, to provide secure design, a bill of materials for components, documented vulnerability handling and the reporting of those that are actively exploited.

What exactly does Trust-Buster provide?

Trust-Buster deploys a sovereign hardware architecture: the Ops-Devices platform.

The solution includes the hardware installed on your site (the Cube for detection and validation, and its satellites for activation and actual power cut-off). We also provide support with enrolling your equipment and your staff, access policy management, and the delivery of signed audit evidence, ready for your compliance checks.

What is Trust-Buster and where is the company based?

Trust-Buster is the access control system for critical equipment built on the sovereign Ops-Devices platform, deployed and operated on site. The company is based in Sophia Antipolis, France.

More technical information about the products?

Visit our technical infrastructure platform from our public website:

https://ops-devices.eu

And where do electronic keys fit into all this?

“That already exists”

For twenty years, electronic key systems have been fitted to machine HMIs to determine which operator level is entitled to which operating mode. It is a proven, industrial device, and we do not claim to replace it.

There are three differences, and they are structural.

They assume a machine that can read a key. An electronic key reader connects to an HMI or a PLC capable of interpreting its value. On equipment with no screen, no identity and no network, which makes up most of an installed base, there is nothing to connect it to. Our enforcement point sits upstream, on the power supply or standard cables, and requires nothing from the machine.

They carry a right, not a proven identity or a time slot. An electronic key is a role token, usually shared, with no automatic expiry, no remote revocation and no check that the intervention was scheduled. It is a usage control, not a chain of authorisation.

They do not produce usable evidence. The trace, when it exists, stays in the PLC or the HMI, in a manufacturer-specific format, unsigned and without an enforceable timestamp.

On a recent machine equipped with an HMI and user management, these systems are often the right choice. Our ground begins where theirs ends: equipment that has nothing to accommodate them.

Connect With Trust-Buster

📍 Sophia-Antipolis — see on map

* required