# Agilicus

Zero Trust Network Architecture

- Site: https://www.agilicus.com/

## When the Foundation Fails: PHP2Shell and the eMudhra Certificate Compromise

- Link: https://www.agilicus.com/when-the-foundation-fails-php2shell-and-the-emudhra-certificate-compromise/
- Published: 2026-10-22T15:57:28+00:00
- Author: Don Bowman

![Shield protecting industrial and technical infrastructure icons including gears,
a server rack, and a power pylon](https://www.agilicus.com/www/d3abff13-shield-protecting-
industrial-and-technical-infrastructure-icons-including.jpg)

## A Week of Broken Foundations

Two incidents landed in the same week, and they rhyme. In systems engineering, we
spend our careers scrutinising application business logic, configuring access control
lists, and reviewing pull requests. We operate on the implicit assumption that the
ground beneath our feet is solid granite. But every so often, the bedrock fractures.

One incident involves a pair of critical flaws in the core PHP runtime, latent in
the codebase since 2012, that permit an unauthenticated HTTP request to execute 
arbitrary code before your application even boots. The other involves [eMudhra](https://emudhra.com/en/post/certificate-security-incident),
a globally trusted Certificate Authority, which generated eight intermediate signing
keys with such pitiful entropy that researchers recovered the private keys from 
public certificates in less than a second. One private key was literally the single-
digit integer 4.

Both failures expose a recurring blind spot in modern infrastructure: we expend 
enormous effort securing the code we write, while blindly trusting the foundations
we inherit.

## PHP2Shell Is Not a Website Problem

When security advisories dropped for the PHP vulnerabilities dubbed PHP2Shell, the
public conversation immediately drifted toward web servers: WordPress installations,
Laravel applications, Drupal portals, and Apache or PHP-FPM daemons running in cloud
environments. The official remediation guidance was predictable: upgrade immediately
to [PHP 8.2.35, 8.3.36, 8.4.27, or 8.5.12](https://www.php.net/archive/2026.php#2026-10-21-5).

If your entire estate consists of modern containerised workloads deployed through
continuous integration pipelines, patching is straightforward. But having spent 
decades building network appliances, industrial equipment interfaces, and high-performance
routers, I look at these advisories through a very different lens. I know where 
PHP actually lives.

PHP is not just on web hosts. It powers the embedded web administration console 
on your branch office router, the configuration interface on a network video recorder,
the local diagnostic panel on a building automation controller, and the embedded
web server inside plant equipment. These devices run on low-cost flash storage, 
often using ancient kernels and stripped-down userlands.

The PHP flaws date back to version 5.4, released in 2012. How many embedded devices
installed in commercial facilities over the past decade still run PHP 5.6 or 7.0?
Nearly all of them. These devices do not possess automated package managers. Many
do not even have an active vendor left to issue a firmware update. As [Cisco Talos noted in its analysis](https://blog.talosintelligence.com/two-critical-rces-in-php-core/),
roughly half of all observed PHP instances in the wild sit on abandoned, end-of-
life releases that will never see an official security patch. The question for your
organisation is not whether you run PHP, but whether you have any physical mechanism
to update the devices that do.

## Execution Before Application Bootstrap

The technical elegance, and terror, of PHP2Shell lies in its timing. The two flaws,
[CVE-2026-109618](https://nvd.nist.gov/vuln/detail/CVE-2026-109618) and [CVE-2026-109619](https://nvd.nist.gov/vuln/detail/CVE-2026-109619),
execute long before your application code receives control.

The first flaw, CVE-2026-109618, exploits out-of-memory error handling. When an 
unauthenticated multipart form submission exhausts memory limits during parsing,
the runtime invokes longjmp() to abandon the request. However, internal request 
structures remain only partially built. Subsequent cleanup routines mistakenly treat
this corrupted, half-initialised memory as valid, creating an exploitable heap corruption
condition. Triggering this condition requires an oversized POST request, often several
megabytes in size.

The second flaw, CVE-2026-109619, is far more dangerous. It leverages PHP’s session
upload progress mechanism, a legacy feature enabled by default that tracks file 
uploads via session variables. When fatal error handling executes, it frees memory
regions while leaving global pointers intact. The very next HTTP request processed
by that identical worker thread inherits the dangling pointer, resulting in a classic
use-after-free scenario. This payload requires no massive transmission: a crafted
request of roughly 450 bytes is sufficient to hijack control flow.

Consider the architectural implication: because the attack executes during initial
request ingestion, none of your application-level defences matter. Your authentication
middleware never runs. Your routing handlers never execute. Your audit logging framework
never records the caller. From the perspective of your monitoring systems, the worker
process simply crashes, restarts, or returns a routine HTTP 200 or 500 error, while
the attacker gains an interactive shell.

## When 384 Bits of Entropy Collapses to the Number 4

While the industry scrambled to inventory PHP runtimes, a second, arguably more 
alarming failure was unravelling across global trust stores. [Mozilla Bugzilla thread 2083449](https://bugzilla.mozilla.org/show_bug.cgi?id=2083449)
published details of a catastrophic key-generation failure by eMudhra, a commercial
Certificate Authority trusted by every major operating system and browser vendor.

During an audited key ceremony conducted on April 30, 2025, eMudhra engineers were
provisioning 24 subordinate Certificate Authorities using hardware security modules.
A scripting error during a wrap and unwrap test routine produced a secondary batch
of test keys. Because the object labels and filenames mirrored the intended production
keys, operators inadvertently selected the test keys to generate official Certificate
Signing Requests, which were duly signed by their root authority and cross-signed
by Sectigo.

These intermediate certificates were configured to use the P-384 elliptic curve.
In standard cryptography, a P-384 private key is a randomly selected 384-bit integer,
a number spanning roughly 116 decimal digits, representing a search space larger
than the total number of atoms in the observable universe. In eMudhra’s blunder,
the entropy collapsed completely. The generated private keys were small integers
under 256: values such as 56, 66, 230, and in one case, the literal number 4.

Using elementary baby-step giant-step algorithms, independent researchers and Mozilla
engineers calculated the private keys directly from the public certificates in less
than a second. For roughly 18 months, eight intermediate Certificate Authorities
with publicly recoverable private keys sat in active production, trusted implicitly
by devices worldwide. Following the disclosure, [Reuters reported that eMudhra has exited](https://www.reuters.com/legal/litigation/indian-firm-emudhra-exits-global-digital-certificates-business-after-2026-10-22/)
the global digital certificate business entirely.

## The Quiet Threat: Code Signing and Authorisation Passports

The standard consumer reaction to a Certificate Authority compromise is to assume
browsers will fix it. Chrome, Firefox, Safari, and Edge pushed rapid updates, leveraging
mechanisms like OneCRL and CRLSets to revoke the intermediate certificates almost
instantaneously. If web browsing were the sole consumer of public key infrastructure,
the crisis would be resolved within 48 hours.

It is not. Of the eight compromised certificates, two were designated for code signing,
and six were general-purpose Transport Layer Security issuers. Both categories inflict
severe architectural damage far beyond the browser.

A code-signing certificate does not prove software is benign; it asserts that a 
specific cryptographic identity signed a binary. Windows operating systems translate
that signature into administrative trust: SmartScreen reputation, Application Control
policies, and AppLocker rules. An attacker holding these intermediate keys could
mint arbitrary code-signing certificates for any corporate identity, sign malware,
and slip seamlessly past enterprise application allow-lists. Because commercial 
code signatures employ cryptographic timestamping, those signed payloads remain 
valid indefinitely, even after the underlying signing certificate expires.

Even more insidious is the role of Transport Layer Security in machine-to-machine
authentication. Modern zero trust architectures frequently deploy mutual Transport
Layer Security across application programming interface gateways, microservice service
meshes, and industrial networks. A client certificate is not merely an encryption
mechanism; it is an authorisation passport. If your industrial supervisory control
and data acquisition system, Internet of Things ingestion hub, or virtual private
network terminates connections using a trust store containing the compromised root
or intermediate certificates, an attacker possessing the private key can forge a
valid client certificate for any identity they choose. They do not need to intercept
your traffic; they authenticate directly as an authorised entity.

## The Trust Stores We Forgot

When you update your desktop browser, you update exactly one trust store. In an 
enterprise operational technology or cloud environment, dozens of unmanaged trust
stores operate silently in the background:

 * The certifi Python package pinned inside hundreds of isolated virtual environments
   and automated machine learning pipelines.
 * The embedded root certificates compiled statically into Go, Rust, and Node.js
   binaries running in container clusters.
 * The Java keystores managing enterprise middleware and message queues.
 * The firmware images on operational technology gateways, network switches, and
   environmental sensors that never contact an Online Certificate Status Protocol
   responder or download a Certificate Revocation List.

Unlike modern web browsers, non-browser clients rarely check revocation status or
enforce Certificate Transparency. As [Cloudflare highlighted in its incident guidance](https://blog.cloudflare.com/emudhra-certificate-compromise/),
well-meaning engineers attempting to resolve connection failures often make matters
worse by disabling certificate verification entirely.

## How We Break the Cycle

Both the PHP2Shell exploit and the eMudhra collapse demonstrate why conventional
perimeter defences inevitably fail. Traditional network architecture relies on two
flawed premises: first, that an internal network segment is secure enough to leave
legacy HTTP interfaces directly reachable; second, that possessing a cryptographic
certificate implies verified authorisation.

When we designed the Agilicus AnyX platform, we explicitly rejected both assumptions.
Our philosophy treats identity and authorisation as the primary security boundary,
entirely decoupling user access from network topology.

Against flaws like PHP2Shell, network-level firewalls and web application firewalls
struggle because they inspect traffic only after a connection is established, and
they frequently fail to parse binary or multipart payloads cleanly. AnyX approaches
the problem from the outside in: unauthenticated traffic never touches the upstream
application runtime. Every inbound request must authenticate against your corporate
identity provider and satisfy multi-factor authentication policies before AnyX routes
a single byte to the destination service.

Furthermore, AnyX enforces granular path-level and method-level authorisation rules.
An external user cannot issue an arbitrary POST request to an undocumented upload
endpoint because AnyX restricts actions strictly to authorised roles and explicit
HTTP paths. If an unauthenticated attacker cannot deliver their 450-byte payload
to the PHP parser, the vulnerability cannot be triggered.

Similarly, against public key infrastructure collapses like eMudhra, AnyX eliminates
blind trust in transport certificates. Instead of allowing client certificates to
grant blanket network access across a traditional virtual private network, AnyX 
validates the human identity behind the request, continuously inspecting context,
device posture, and role assignments.

## The Monday Morning Audit

To protect your operations against the fallout from these two foundation failures,
take the following concrete steps:

 * **Isolate legacy web consoles:** Inventory every operational technology device,
   router, print server, and network camera. Place them behind authenticated access
   boundaries so unauthenticated network traffic cannot reach their embedded web
   runtimes. Review our tactical guide on [reducing exposure to the PHP2Shell vulnerability family](https://www.agilicus.com/faq/how-do-i-reduce-exposure-to-the-php2shell-remote-code-execution-family/).
 * **Audit non-browser trust bundles:** Inspect your container base images, Python
   virtual environments (certifi), and Java runtimes to ensure the eight compromised
   eMudhra intermediate keys and associated roots are explicitly purged.
 * **Enforce granular path authorisation:** Restrict HTTP POST, PUT, and DELETE 
   methods across all web services to authenticated roles, and eliminate wildcards
   in your reverse proxy configurations. See our documentation on [authorisation rules](https://www.agilicus.com/anyx-guide/authorisation-rules/)
   and [web application security](https://www.agilicus.com/anyx-guide/web-application-security/).
 * **Re-evaluate application allow-lists:** Audit your Windows Defender Application
   Control and AppLocker policies to ensure binary execution trust is grounded in
   cryptographic hashes rather than publisher certificates alone.

We are accustomed to securing the applications we write while taking the foundation
for granted. But when the runtime executes attacks before your code runs, and when
a global certificate authority signs keys with single-digit entropy, the foundation
has failed. The organisations that survive these shifts are not the ones with the
fastest patching cycles, but the ones whose architecture never trusted the foundation
in the first place.

## The Air Gap is Dead: Water Utilities Must Embrace Zero Trust for OT Security

- Link: https://www.agilicus.com/the-air-gap-is-dead-water-utilities-must-embrace-zero-trust-for-ot-security/
- Published: 2026-07-14T23:59:02+00:00
- Author: Don Bowman

![Industrial infrastructure connected to a secure cloud with biometric authentication
and security protection](https://www.agilicus.com/www/c6cea989-industrial-infrastructure-
connected-to-a-secure-cloud-with-biometric.jpg)

I recently had the privilege of speaking at the Water AI 2026 summit in Anaheim,
where I shared my thoughts on a critical intersection of modern infrastructure: 
Operational Technology (OT) security and the reality of the industrial cloud. As
water utilities continue their march into Industry 4.0 — an era defined by cyber-
physical systems, distributed control, and edge AI — the way we defend our critical
infrastructure has to fundamentally change.

For decades, the water industry has operated on long timelines. A plant built today
is expected to run for 25 years. Capital expenditure (CapEx) rules the day, and 
once equipment is bolted to the floor, it stays there. But while the physical pumps
and agitators remain constant, the digital landscape around them has evolved at 
a breakneck pace. We are no longer living in the era of isolated, local-only SCADA
systems. We are operating cyber-physical systems where digital commands directly
manipulate heavy machinery.

As we connect these systems to leverage big data, remote operations, and AI, we 
must confront an uncomfortable truth: the traditional security models we’ve relied
on are failing us.

### The Myth of the Air Gap

Historically, OT security leaned heavily on the Purdue Model and the concept of 
the “air gap.” The philosophy was simple: outside is infinitely bad, inside is infinitely
good. It’s like the human body relying entirely on skin to keep infections out, 
but lacking white blood cells to fight an infection if the skin gets scratched.

In reality, the air gap is dead.

I’ve walked into multi-million-gallon-per-day wastewater plants where facility managers
swore their networks had absolutely zero outside connectivity. Two minutes later,
I spotted 700MHz LTE antennas bolted to a cabinet — installed by a vendor for remote
maintenance without the plant operator’s knowledge. Between vendor shadow IT, dial-
in modems, and undocumented internet connections, your environment is likely far
more connected than you realise.

### The Cold, Hard Facts of OT Attacks

When we look at the data from the Cybersecurity and Infrastructure Security Agency(
CISA) Risk & Vulnerability Assessments, the narrative of sophisticated, movie-style
hacking falls apart. Over 42% of successful attacks use valid, existing accounts.
Another 26% rely on spear-phishing.

In total, roughly 80% of successful attacks exploit user authentication and identity
weaknesses. Attackers aren’t breaking through your firewalls with zero-day exploits;
they are simply logging in. They find a shared credential like “Siemens_Admin,” 
exploit a legacy VPN that provides broad access to the entire plant network, and
move laterally until they reach your critical PLCs and HMIs.

### Moving to a Zero Trust Framework

Security vendors often treat “Zero Trust” as a product you can buy off a shelf, 
but it’s actually a framework. It boils down to three core pillars:

Identity (**Who**): Every user must be individually known and authenticated. We 
must eliminate shared accounts. Whether it’s an internal operator or a third-party
vendor, identity must be tied to a specific human.

Authorisation (**What**): Access must be governed by the principle of least privilege.
An HVAC contractor shouldn’t have network-wide access to your water treatment PLCs.
They should only have access to the specific application they need, and perhaps 
only during specific hours.

Access (**How**): We need to kill the corporate VPN. VPNs place users directly onto
the network. Instead, we should use Identity-Aware Proxies that route users only
to the specific destination resources they are authorised for, entirely decoupled
from the underlying network topology.

### A Non-Zero-Sum Game

In the IT world, there is a long-standing belief that security and productivity 
are a zero-sum game — if you increase security, you make a system harder to use.
In the OT space, Zero Trust is actually a non-zero-sum game. You can increase security,
boost productivity, and lower costs simultaneously.

Think about your task workers and external experts. As the workforce ages out and
the skills gap widens, you will increasingly rely on remote subject matter experts.
By implementing Zero Trust, these experts can access the resources they need from
any device, on any network, without installing clunky VPN clients. By adopting modern
authentication like Passkeys (which use local biometrics like your fingerprint or
FaceID), you can eliminate passwords entirely, defeating phishing while making the
login process seamless.

### Implementation: Hours, Not Months

The most common objection I hear is that overhauling plant security is too expensive
and complex. But modernising your OT security doesn’t require ripping and replacing
your legacy infrastructure.

With an Identity-Aware Proxy, you can bolt modern OpenID Connect authentication 
onto 25-year-old PLCs and HMIs without changing a single line of config on the plant
floor. Furthermore, by utilising an outbound-only architecture, you can close all
inbound firewall ports. The plant establishes a secure, outbound tunnel to the cloud,
dramatically reducing your attack surface while still allowing authorised users 
to connect inward securely.

You don’t have to fix everything on day one. You can deploy this incrementally —
one user, one application, one vendor at a time.

The industrial revolution of water utilities is already here. The machines are talking
to the cloud, and remote access is a necessity, not a luxury. By embracing a Zero
Trust framework, we can build a defence-in-depth posture that protects our critical
water infrastructure without getting in the way of the people working hard to keep
it running.

## The Epitome of Absolute Trust: Why Legacy Virtual Private Networks Are a Liability

- Link: https://www.agilicus.com/the-epitome-of-absolute-trust-why-legacy-virtual-private-networks-are-a-liability/
- Published: 2026-06-09T01:48:23+00:00
- Author: Don Bowman

![Abstract digital illustration representing the shift from a vulnerable legacy 
virtual private network to modern identity-based security.](https://www.agilicus.
com/www/6b50e998-z.png)

Recently, a [critical vulnerability](https://www.cyber.gc.ca/en/alerts-advisories/check-point-security-advisory-av26-559)
carrying a severity score of 9.3 was announced for Check Point remote access deployments.
The flaw centres on the deprecated Internet Key Exchange version 1 protocol. This
is a standard that has been obsolete for years. Yet, it is still sitting on the 
perimeter of countless organisations, acting as an open door. Threat actors, specifically
affiliates of the Qilin ransomware group, are already exploiting it in the wild.

This is not an isolated incident. The [2025 VPN Risk Report](https://www.zscaler.com/learn/2025-vpn-risk-report)
highlights that virtual private networks remain the top exploited area for enterprise
infrastructure. If we look at the Known Exploited Vulnerabilities catalog maintained
by the [Cybersecurity and Infrastructure Security Agency](https://www.cisa.gov/known-exploited-vulnerabilities-catalog),
it is heavily populated with perimeter security appliances. Just this year, we have
seen major warnings issued for Palo Alto firewalls and Fortinet systems. The [Canadian Centre for Cyber Security](https://www.cyber.gc.ca/en/alerts-advisories/exploitation-fortinet-fortios-vulnerabilities-cisa-fbi)
recently issued stark warnings regarding active exploitation of these very appliances.

We have discussed previously how the [time to exploit approaches zero](https://www.agilicus.com/time-to-exploit-approaches-zero/).
Once a vulnerability is public, automated scanners find your exposed endpoints in
minutes. The fundamental issue is that a virtual private network is the epitome 
of absolute trust. It is the exact opposite of a zero trust architecture.

## The Operational Nightmare of Absolute Trust

When you provision a virtual private network connection, you are effectively running
a long, invisible ethernet cable right into your deep infrastructure. It grants 
broad network access rather than narrow, application-specific access.

Compounding this architectural flaw is the operational nightmare. These software
clients are set up and forgotten. They reside on the laptops of third-party vendors,
contractors, and partners. They span multiple management domains where your internal
remote monitoring and management tools cannot reach. When a zero-day vulnerability
drops, you face an impossible task. You cannot force an update on a machine you 
do not control. You are left relying on external partners to patch their clients,
while your perimeter remains critically exposed.

## Modernising Access with Strong Identity

The industry must stop attempting to patch legacy network bridges and start modernising
access. The answer lies in strong, simple identity. We need to transition away from
network-level trust and move to modern, web-based standards utilising HTTPS and 
TLS 1.3.

By deploying an Identity-Aware Proxy, you remove the network layer from the equation
entirely. Users authenticate with strong multi-factor authentication and single 
sign-on. They are granted access to specific applications rather than the entire
corporate network. If a partner device is compromised, the blast radius is contained
strictly to the authorised application, rather than providing a pivot point into
your core infrastructure.

It is time to accept that perimeter-based network access is a liability. Transitioning
to identity-first security is the only pragmatic way to protect your resources in
an era where the time to exploit is practically non-existent.

## Defence in Depth: Zero Trust is a Critical Layer, Not a Panacea

- Link: https://www.agilicus.com/defence-in-depth-zero-trust-is-a-critical-layer-not-a-panacea/
- Published: 2026-06-05T00:01:23+00:00
- Author: Don Bowman

![Defence in Depth and Zero Trust Strategy](https://www.agilicus.com/www/7d0ce2c6-
z.png)

In the world of cyber security, the overall strategy I subscribe to is “Defence 
in Depth.” In a nutshell, this means creating a multi-layered system where each 
layer addresses non-correlated risks. You must assume that each layer is non-absolute
in its protection: eventually, a layer will be breached. By layering your security,
you are effectively buying time.

The classic medieval example is the castle built on a hill, surrounded by a moat,
revetments, and a keep. Because it is on a hill, you see the attackers coming. The
moat and walls slow them down. Each obstacle provides the defenders with time to
react, reposition, and respond.

Zero trust is a vital part of this defence in depth strategy, but it is not an absolute
solution. It is comprised of three primary elements:

 * **The WHO problem**: Authentication. Who are you?
 * **The WHAT problem**: Authorisation. What are you allowed to do?
 * **The HOW problem**: Access. How do I get you to that resource?

Zero trust means solving for who, what, and how on a per-person basis (never using
shared accounts), using fine-grained permissions rather than all-or-nothing access,
and doing so per-resource rather than for an entire network at once, like a traditional
virtual private network. It also means placing zero trust in how you attach to the
network. You must prove who you are for each specific resource: you cannot just 
connect to a virtual private network or plug in an ethernet cable and assume you
are safe.

## AI as a Risk Amplifier

Now, let’s consider the risks associated with artificial intelligence. While AI 
has its own unique risks, it is primarily an amplifier of existing ones. Risk is
a simple equation: Threat multiplied by Vulnerability multiplied by Outcome. The
primary thing artificial intelligence does is decrease the cost associated with 
exploiting a vulnerability.

When we look at the “Threat,” we ask: who is the attacker? Is it a nation-state 
actor, or a tired systems administrator who accidentally rebooted the wrong server?
The “Vulnerability” might be a known software flaw or a misconfiguration. The “Outcome”
is the bad thing that happens: does it affect a municipal fresh water supply, or
does a printer simply run out of toner?

A recent [infographic from the Cybersecurity and Infrastructure Security Agency (CISA)](https://www.cisa.gov/sites/default/files/2024-09/InfographicFY23RVA508.pdf)
provides excellent context here. It maps the actual attacks from 2023 to the MITRE
ATT&CK framework. The primary takeaway is “credentials.” Valid accounts, spearphishing
to obtain valid accounts, and brute force guessing of accounts comprise approximately
80 per cent of successful attacks.

The old Italian philosopher Vilfredo Pareto once described an 80:20 rule, and it
holds true in cyber security today.

## Our Philosophy of Zero Trust

Because credentials are the primary target, a large part of our philosophy involves
three core principles:

 1. **Remove the need for shared accounts**: Every human should sign in with their 
    single, natural, native credentials. Whether you are an internal employee or an
    external contractor, you should use single sign-on with properly managed identity
    systems that implement heuristics around password guessing, strength, etc.
 2. **Increase authentication strength**: We advocate for multi-factor authentication,
    which requires any two or more of: something I have, something I know, or something
    I am. This goes back to the non-correlated risks mentioned earlier. The method 
    someone uses to guess a password is entirely different from how they might steal
    a physical security key. These risks are divisive rather than additive.
 3. **Remove lateral traversal**: I call this reducing the “blast radius.” If a team
    member has their signed-in phone stolen at lunch, what can the thief do? In a zero
    trust environment, they can only do what that specific user is authorised to do,
    within the limits of the access token recently established.

Is zero trust a panacea? Certainly not. Belts and suspenders exist for a reason.
Submarines have internal watertight doors for a reason. It is part of a balanced
breakfast of security measures.

## The Marketing of AI Danger

I should note that much of the current “mythos” surrounding the danger of artificial
intelligence is a brilliant marketing campaign. Claims that a model is “too dangerous
to release” create massive demand. While models from OpenAI, Google, and others 
are amazing, the marketing PR teams are genius at framing risk as a feature.

If I were to attack an organisation like yours using AI, I would focus on human 
engineering. Business email compromise via AI-generated emails can be incredibly
convincing. For example, an attacker might scrape information from open-source intelligence
to send a message like: “Hey, it is Joe from the XYZ team. Remember that company
picnic last week? I forgot my laptop in the office and need the RFP responses: could
you please send them to my personal account?”

The human “lizard brain” starts trusting very fast when presented with information
it assumes would only be known by a real party (often called the “facebook scam”).
AI also reduces the cost of brute force attacks, such as guessing wifi passwords
or scanning for vulnerabilities. In short, AI changes the bar for attacking you.
A small municipality might not be a bank full of intellectual property, but with
AI, a lazy attacker can cast a much wider net.

## The Asymmetric Cost of Defence

Does AI render existing security principles invalid? No. Does it increase the already
asymmetric cost of defending versus attacking? Absolutely. The blue team (defenders)
needs to be right 24 hours a day, 365 days a year. The red team (attackers) only
needs to be right once.

This is why defence in depth is so critical. Give yourself visibility, slow the 
attackers down, and give yourself time. Keep the section doors closed in the submarine.

As nation-state cyber saber-rattling has ratcheted up, the target has changed. A
wastewater plant in a mid-size municipality might not be a dollar-rich target for
criminals, but it is an embarrassing target to hit for geo-political reasons. We
have seen this in [Aliquippa](https://www.agilicus.com/avoid-exploitation-of-unitronics-plcs-used-in-public-water-systems/),
Pennsylvania, and [Muleshoe](https://www.agilicus.com/inbound-hmi-cyber-army-of-russia-targeting-us-water-facilities/),
Texas. In the risk equation, the “Threat” has gone up even if the “Vulnerabilities”
remain the same.

I am also concerned that we are crossing the singularity of zero-day disclosure.
In 2020, the time from discovery to disclosure was 745 days. In 2025, it was 44 
days. In a perfect world, a vendor has only 44 days to solve, test, package, and
deliver a patch. This is approaching the 10-day range. This means that mitigating
measures, such as an identity-aware web application firewall, are becoming even 
more important than the “patch early, patch often” mantra.

With our system, combining an identity-aware web application firewall with best-
in-class identity providers, I am confident it remains a relevant and necessary 
part of your defence in depth.

**Related Content:**

 * [Practical Steps for Wastewater Treatment Plants](https://www.agilicus.com/advice-avalanche-practical-steps-for-wastewater-treatment-plants/)
 * [Understanding Multi-Factor Authentication](https://www.agilicus.com/i-am-i-have-i-know-multi-factor-authentication/)
 * [Historical Context: Early Security Strategy Video](https://youtu.be/ClQNbiVWW4s)
 * [AI-Powered Cyber Threats and Critical Infrastructure](https://www.agilicus.com/ai-powered-cyber-threats-protecting-your-critical-infrastructure/)

## Securing the Flow: GAO Highlights Persistent Cyber Threats to Water and Wastewater Systems

- Link: https://www.agilicus.com/securing-the-flow-gao-highlights-persistent-cyber-threats-to-water-and-wastewater-systems/
- Published: 2026-05-22T13:53:46+00:00
- Author: Don Bowman

![Water Infrastructure Cybersecurity Zero Trust](https://www.agilicus.com/www/d55b6032-
9k.png)

On May 21, 2026, the United States Government Accountability Office (GAO) released
a critical report, [GAO-26-109159](https://www.gao.gov/products/gao-26-109159), 
addressing the urgent and persistent cybersecurity threats facing the nation’s water
and wastewater sector. This testimony, provided to the House Committee on Science,
Space, and Technology, underscores a stark reality: nearly 170,000 systems are increasingly
vulnerable to cyberattacks from state-sponsored hackers and criminal groups. (Read
the [highlights](https://www.gao.gov/assets/gao-26-109159-highlights.pdf) or the
[full report](https://www.gao.gov/assets/gao-26-109159.pdf)).

For municipal leaders and plant managers, the GAO report is more than a warning;
it is a call to action to **modernise** how we protect our most essential critical
infrastructure.

## The Rising Tide of Cyber Threats

The water and wastewater sector is currently facing a growing number of sophisticated
cyberattacks. These threats are no longer merely theoretical. Adversaries like the
[Cyber Army of Russia](https://www.agilicus.com/inbound-hmi-cyber-army-of-russia-targeting-us-water-facilities/)
are actively targeting operational technology systems, including the human-machine
interfaces (HMIs) and programmable logic controllers (PLCs) that control physical
infrastructure like pumps, valves, and chemical levels.

The GAO identifies several factors contributing to this heightened risk profile.
First, the convergence of operational technology with internet-enabled devices has
significantly expanded the attack surface. Second, the sector faces persistent resource
constraints, where limited financial budgets often force a choice between meeting
basic water safety regulations and investing in cybersecurity. Finally, workforce
shortages and the difficulty of updating aging infrastructure make it challenging
to maintain a robust security posture across diverse and often geographically dispersed
systems.

## The Myth of the Air Gap and the Failure of Perimeter Security

Historically, the water industry relied on the [myth of the air gap](https://www.agilicus.com/manitoba-water-wastewater-association-why-water-facilities-must-embrace-zero-trust-in-industry-4-0/)
to ensure security, assuming that operational technology systems were physically
disconnected from the outside world. In the era of Industry 4.0, this air gap is
dead. Modern plants require remote support from vendors, data exports for regulatory
compliance, and seamless connectivity for operational efficiency.

To facilitate this access, many facilities have relied on traditional virtual private
networks. However, as the GAO report suggests, these perimeter-based defences are
no longer sufficient. A virtual private network functions like an elongated Ethernet
cable, granting broad network access once a user is authenticated. If a single credential
is compromised, an attacker gains _the keys to the castle_, allowing them to move
laterally through the network to discover and manipulate critical devices.

## Agilicus AnyX: Building Defence in Depth through Zero Trust

The Environmental Protection Agency (EPA) and other federal partners are working
to address these risks, but the GAO notes that significant gaps in authority and
strategy remain. While we wait for federal mandates to catch up, facilities can 
take immediate, proactive steps to implement a true **defence in depth** strategy
using zero trust architecture.

Agilicus AnyX technology is designed to solve these exact challenges by shifting
security away from the network perimeter and focusing on the identity of the user
and the specific resource they need to access.

### Unified Identity and Multi-factor Authentication

The first pillar of zero trust is ensuring that every user is individually known.
Agilicus AnyX enables unified authentication, allowing operators and contractors
to sign in with their native corporate credentials through a single sign-on experience.
By enforcing strong multi-factor authentication, we eliminate the risks associated
with shared administrative accounts or default passwords on exposed controllers.

### Precise Authorisation and Lateral Movement Prevention

Once identity is verified, access must be strictly limited. Agilicus allows for 
granular authorisation, ensuring that a vendor or technician only sees the specific
human-machine interface or programmable logic controller they are approved to maintain.
This _pairwise_ connection effectively stops lateral traversal; even if a user device
is compromised, the attacker cannot pivot to other parts of the critical infrastructure.

### Absolute Segmentation and Inbound Port Removal

A key recommendation for the water sector is to ensure there are zero inbound open
ports. Agilicus AnyX utilizes outbound-only connections to an identity-aware proxy,
making your infrastructure completely invisible to [public-endpoint scanner](https://www.agilicus.com/cisa-cyber-scan-water/)
tools like Shodan or Nmap. This achieves a level of industrial micro-segmentation
that traditional firewalls and virtual private networks simply cannot provide.

## Conclusion

The persistent threats highlighted by the GAO report require a fundamental shift
in how we approach cybersecurity for water and wastewater systems. We cannot afford
to wait for the next incident response warning to upgrade our defences.

By adopting a zero trust architecture, municipal water facilities can enhance their
security posture, protect public health, and ensure operational continuity without
sacrificing the efficiency of remote access. It is time to move beyond the myth 
of the air gap and embrace a pragmatic, results-oriented approach to critical infrastructure
protection.

## NERC CIP Compliance for Small Independent Power Producers: A Pragmatic Approach

- Link: https://www.agilicus.com/nerc-cip-compliance-for-small-independent-power-producers-a-pragmatic-approach/
- Published: 2026-05-18T10:48:00+00:00
- Author: Don Bowman

![NERC CIP Compliance for Remote Power Sites](https://www.agilicus.com/www/4c77c9c8-
2q.png)

Small independent power producers face a unique set of challenges when it comes 
to regulatory compliance. The Bulk Electric System is moving towards stricter enforcement
of cybersecurity standards, and small facilities often lack the massive IT teams
or budgets of larger utilities. With the upcoming enforcement of [CIP-003-9](https://www.nerc.com/pa/Stand/Reliability%20Standards/CIP-003-9.pdf),
the stakes have never been higher. A single violation can lead to significant fines.
For the pragmatic plant manager, the goal is not just to check a box, but to build
a resilient operation that maintains uptime while meeting every requirement of the
North American Electric Reliability Corporation.

## The Shift to CIP-003-9 and Vendor Remote Access

The new standard introduces critical sections regarding vendor electronic remote
access security controls. Specifically, sections 6.1 and 6.2 require organisations
to determine how vendors access systems and ensure that access is disabled immediately
when no longer needed. Many small power sites still rely on [legacy tools](https://www.agilicus.com/white-papers/nerc-cip-003-9-legacy-remote-access-tools-are-not-sufficient/)
like virtual private networks or jump boxes. While these were once the industry 
standard, they are increasingly seen as a liability. They often lack the granular
visibility required to prove compliance during an audit. If you cannot identify 
exactly who accessed a specific controller at 2:00 PM on a Tuesday, you are at risk.

## Why Multi-Factor Authentication is Non-Negotiable

Multi-factor authentication is one of the most impactful safeguards an independent
power producer can implement. By requiring two or more forms of verification, such
as a password and a physical token or biometric, you reduce the risk of credential
theft significantly. The [Cybersecurity and Infrastructure Security Agency](https://www.cisa.gov/resources-tools/resources/zero-trust-maturity-model)
notes that this simple step can prevent the vast majority of common cyber attacks.
For low-impact assets, this is no longer just a suggestion. It is becoming a core
part of the roadmap for the North American Electric Reliability Corporation. Implementing
a tool that enforces multi-factor authentication across all remote sessions is a
pragmatic first step towards a more secure site.

## The Pitfalls of Shared Credentials

Many legacy systems rely on shared vendor accounts. This makes it impossible to 
distinguish between different individuals from the same service provider. Single
sign-on solves this by allowing vendors to use their own company credentials. When
a vendor technician leaves their organisation, their access is automatically revoked.
This provides a built-in kill switch that directly addresses the requirements of
[CIP-003-9](https://www.nerc.com/pa/Stand/Reliability%20Standards/CIP-003-9.pdf)
section 6.2. Without single sign-on, you are left managing a revolving door of passwords
and hoping that your vendors inform you when an employee departs. This is not a 
sustainable or compliant way to operate a modern power facility.

## Least Privilege and Granular Audit Trails

The principle of least privilege ensures that a user only has access to the specific
resources they need to perform their job. If a technician only needs to monitor 
a sensor, they should not have write access to the main controller. Zero trust architectures,
like those provided by Agilicus, operate at the application layer rather than the
network layer. This allows for incredibly granular [audit logs](https://www.agilicus.com/white-papers/nerc-cip-003-9-automating-access-control-and-authorisation-for-remote-power-sites/).
You can see who accessed what, when they did it, and what actions they took. This
centralized data simplifies the evidence collection process for any compliance manager.
Instead of hunting through disparate network logs, you have a single source of truth
that is ready for any audit.

## Conclusion: Moving Beyond the Perimeter

Relying on perimeter-based security is a strategy of the past. As we move into the
era of Industry 4.0, small independent power producers must adopt tools that offer
more than just a gate at the edge of the network. By implementing a zero trust platform,
you ensure that every access request is authenticated and authorised individually.
This not only meets the current requirements of NERC [CIP-003-9](https://www.nerc.com/pa/Stand/Reliability%20Standards/CIP-003-9.pdf)
but also prepares your facility for future regulatory shifts. It is a pragmatic,
cost-effective way to secure your assets and focus on what you do best: generating
reliable power for the grid.

## Cyber Security Is Physical Safety: Lessons From a Thermal Runaway

- Link: https://www.agilicus.com/cyber-security-is-physical-safety-lessons-from-a-thermal-runaway/
- Published: 2026-05-17T02:32:31+00:00
- Author: Don Bowman

![Cyber-Physical Safety: EV Charger Thermal Runaway Warning](https://www.agilicus.
com/www/5a7bf80e-2q.png)

On Oct. 11, 2025, a fast-food restaurant in New Westminster, B.C., was evacuated
as toxic emissions and a raging fire consumed a nearby electric vehicle charging
station. The incident destroyed the unit, damaged adjacent infrastructure and serves
as a stark warning for the industrial sector. This was not a random equipment failure.
It was a cyber-physical event where software logic was allowed to override physical
reality.

## Anatomy of a Remote Error

According to Technical Safety BC report II-1977850-2025, the fire was caused by 
a human error during remote maintenance. A technician based in the United States
was performing diagnostics on the cooling system of a battery-integrated fast charger.
During this process, the technician inadvertently activated the battery heater instead
of the cooling fans.

In many systems, this would be a minor configuration error. However, the charger
was in a “diagnostic mode” that intentionally suppressed critical safety measures.
These safety protocols, designed to shut down the system if temperatures became 
unsafe, were effectively bypassed. For 26 hours, the battery heater ran continuously
while the safety systems remained silent. The battery temperature climbed to 82 
C, triggering a thermal runaway event.

## The Software Safety Trap

The incident highlights a dangerous trend in industrial design: the over-reliance
on software-based safety. We often treat software as a universal fail-safe, assuming
that a few lines of code can prevent disaster. But as this incident proves, software
is inherently malleable. It can be toggled off for testing, bypassed by diagnostic
modes or subverted by malicious actors.

If a well-meaning technician can accidentally bypass safety protocols and cause 
a fire, imagine what a motivated attacker could achieve. We are entering an era 
where malware can be a kinetic weapon. A malicious actor with access to a remote
gateway could activate heaters across a regional network of chargers or override
pressure valves in a gas distribution system. When code controls the physical world,
a software vulnerability becomes a public safety hazard.

## The Pragmatic Solution: Hardware Interlocks

To move toward true cyber resilience, we must return to a fundamental engineering
principle: critical safety limits must be enforced by local hardware.

If a software failure or an administrative override can lead to a fire or explosion,
that system is fundamentally insecure. Physical “limit switches” and hardware-based
thermal cut-offs must exist outside the reach of any software diagnostic mode. These
interlocks should be impossible to override remotely, ensuring that no matter what
happens at the control layer, the physical equipment remains within its safe operating
envelope.

## Securing the Remote Gateway

Remote maintenance is a necessity for modern infrastructure, but it requires a more
robust approach than simply lowering the shields. The reliance on “silent email 
alerts” observed in the New Westminster incident is a failure of communication and
accountability.

Instead of broad diagnostic modes that suppress safety, organisations need strong,
identity-based safeguards. Every remote session should be governed by a zero trust
architecture that validates the user’s identity and restricts their actions to the
bare minimum required for the task.

 * **Unified Authentication:** Technicians should use their native corporate credentials
   through single sign-on. This ensures that access is automatically revoked when
   a technician leaves their role.
 * **Precise Authorisation:** Access should be granular. A technician diagnosing
   an HVAC system should not have the ability to override battery thermal protections.
 * **Real-Time Accountability:** Critical alerts must be more than silent emails.
   They require immediate, identity-verified acknowledgement and escalation.

## Conclusion: A New Security Mandate

Cyber security is no longer just about protecting data or preventing downtime; it
is about physical safety. The fire in New Westminster is a call to action for every
industrial operator. It is time to audit our systems for software-only safety traps
and re-implement the hardware interlocks that keep our communities safe.

Don’t let code be the only thing between your infrastructure and a thermal runaway.
True safety starts when the software stops and the hardware takes over.

## Water utility cybersecurity: The nearly £1 million lesson from South Staffordshire Water

- Link: https://www.agilicus.com/water-utility-cybersecurity-the-nearly-1-million-lesson-from-south-staffordshire-water/
- Published: 2026-05-16T12:57:00+00:00
- Author: aman bhogal

![Industrial water facility with digital security overlay representing Zero Trust
for critical infrastructure security.](https://www.agilicus.com/www/c32e02b9-z.png)

# Water utility cybersecurity: The nearly £1 million lesson from South Staffordshire Water

The Information Commissioner’s Office (ICO) recently fined South Staffordshire Water
£963,900. The penalty follows a major data breach where attackers lurked in internal
systems for over a year before anyone noticed. While the breach was discovered in
August 2022, the investigation found attackers had been inside since 2021. For water
utilities and essential services, the lesson is simple: the “castle and moat” security
model is dead. Traditional perimeters just don’t work in a modern threat environment.

## Why the perimeter failed

The South Staffordshire Water breach highlights a fatal flaw in traditional security
thinking. Most organizations still dump their budgets into firewalls, assuming the
internal network is a safe zone. That assumption collapses the moment an attacker
steals one set of credentials or an employee clicks a phishing link. Once inside
the perimeter, attackers find themselves in a high-trust zone with almost no barriers
between them and critical operational assets.

Because South Staffordshire Water didn’t have robust internal monitoring, the breach
went undetected for two years. Traditional security looks at traffic at the gateway,
but it misses what users do once they’re authenticated. If an attacker uses a stolen
password for a virtual private network, they look like a trusted team member. Without
identity-aware access controls that verify the user, the device, and the context
of every single request, your perimeter is an open door. Security has to follow 
the identity of the user and the specific application, regardless of their location
on the network. For a deeper look at modernizing your defences, read our [Industrial Cyber Security Best Practices Guide](https://www.agilicus.com/resources/industrial-cyber-security-best-practices/).

## Identity is the new perimeter

The ICO noted that South Staffordshire Water failed to use multi-factor authentication
for remote access. This is a massive oversight for a critical public service. While
multi-factor authentication is a basic requirement today, many organizations still
struggle to deploy it across legacy systems and operational technology. The Agilicus
Zero Trust platform solves this without the usual deployment headaches.

By replacing passwords and virtual private networks with identity-aware access, 
Agilicus ensures that stolen credentials aren’t enough to breach a system. When 
a user logs in, we verify their identity through single sign-on with multiple factors.
This happens at the application layer, not the network layer. If a phishing attack
compromises a password, the attacker is still blocked because they can’t provide
the secondary factor. We effectively remove the risk of credential theft becoming
a single point of failure.

Our platform also enforces granular access. Unlike a virtual private network that
grants a user access to an entire network segment, Agilicus only allows access to
the specific application needed for a task. This “least privilege” approach is critical
for infrastructure security. Even if one system is compromised, the attacker is 
trapped in that application and can’t move laterally to other assets.

## The five per cent visibility trap

Most water utilities only monitor about five per cent of their internal network 
traffic. Traditional tools are overwhelmed by the volume of data from industrial
systems. But as South Staffordshire Water found, what you don’t see can cost millions
in fines and remediation. If attackers can lurk for years, you have a visibility
problem.

We change this with a 100 per cent audit trail at the application layer. Agilicus
records every access request, login, and denial in a searchable format. This turns
monitoring from a passive checkbox into an active defence. Security teams can see
exactly who accessed what and what they did. This level of auditability is now a
regulatory necessity for any organization managing critical services. To identify
your own gaps, our [Cyber Security Assessment](https://www.agilicus.com/cyber-security-assessment/)
provides a roadmap for improvement. You can’t fix what you can’t see.

## Legacy infrastructure doesn’t need to be a security hole

Water utilities rely on legacy industrial control systems designed decades ago. 
They weren’t meant for the internet and often lack multi-factor authentication. 
Many operators fear that modern security means expensive hardware or disruptive 
downtime. This fear often delays critical improvements.

The Agilicus approach handles this “legacy headache” by wrapping a security layer
around these older systems. It doesn’t require code changes or agents on sensitive
equipment. You can implement Zero Trust for critical infrastructure today using 
the systems you already have. This non-disruptive approach lets utilities meet modern
standards without risking operational uptime. It’s a pragmatic way to bring twentieth-
century infrastructure into a modern security environment.

## Conclusion: A pragmatic approach to security

The fine against South Staffordshire Water is a wake-up call for the entire utility
industry. It’s a reminder that doing nothing costs far more than implementing modern
controls. But effective security isn’t about building higher walls. It requires 
a fundamental shift in how we think about access, identity, and trust.

Every request must be verified, every action audited, and every user granted only
the minimum access they need. At Agilicus, we prioritize business outcomes and reliability.
By focusing on identity-aware access and application-level visibility, utilities
can avoid the failures that led to the South Staffordshire breach. It’s time to 
move beyond the perimeter and embrace Zero Trust for modern critical infrastructure.

## Fortifying Critical Infrastructure: The Shift from Defensible to Isolated

- Link: https://www.agilicus.com/fortifying-critical-infrastructure-the-shift-from-defensible-to-isolated/
- Published: 2026-05-11T09:56:00+00:00
- Author: Don Bowman

![Critical Infrastructure Fortify Architecture](https://www.agilicus.com/www/be9fa13e-
9k.png)

In the quiet corners of our utility rooms and control centres, a strategic shift
is underway. For decades, we have relied on the notion of a defensible perimeter,
a digital wall intended to keep the bad actors out while allowing engineers and 
technicians to operate from within. But as the geopolitical landscape shifts and
sophisticated nation-state actors like **[Volt Typhoon](https://www.cisa.gov/news-events/cybersecurity-advisories/aa24-038a)**
preposition themselves within our vital networks, that wall is more of a liability
than a safeguard. The Cybersecurity and Infrastructure Security Agency (CISA) has
responded with the **[Critical Infrastructure Fortify](https://www.cisa.gov/topics/industrial-control-systems/ci-fortify)**
initiative. This isn’t just another set of guidelines; it is a call to action that
urges operators to move beyond reactive defence and embrace a posture of proactive
isolation. This initiative is built on three sobering assumptions: our connectivity
will be unreliable, our networks are likely already compromised, and we must be 
prepared to operate in an isolated state for weeks or even months. This is a fundamental
shift in architecture that requires us to rethink the nature of secure connectivity.

The first pillar of this new reality is accepting **unreliable connectivity**. In
a geopolitical crisis, the telecommunications and internet services we take for 
granted may become unavailable or untrusted. The initiative explicitly warns that
threat actors are targeting the software and hardware that facilitate our connections
to third-party vendors and business networks. Traditional security models, which
depend on constant communication with external authentication servers and cloud-
hosted management consoles, will fail when those links are severed. Operators must
now identify the vital **[Operational Technology](https://www.cisa.gov/resources-tools/resources/secure-connectivity-principles-operational-technology-ot)**
required to maintain minimum service delivery targets while completely disconnected
from the broader internet. This means moving away from a ‘defence-in-depth’ model
that is porous and reactive toward a ‘resilience-by-design’ model where systems 
are architected to be self-sufficient. By proactively planning for isolation, we
can cut off the command-and-control channels used by adversaries while ensuring 
that our water, energy, and transportation systems continue to serve the public,
even in a degraded state.

However, the greatest obstacle to this isolation is often the very tool we once 
considered a security standard: the **Virtual Private Network** (VPN). VPNs are 
anchors that prevent true resilience in critical infrastructure. By design, a VPN
creates a persistent, trusted path into the heart of a network, often granting broad
access to an entire segment. If an adversary compromises a technician’s credentials
or exploits a vulnerability in the VPN software itself, they have a literal highway
for lateral movement across the environment. Furthermore, VPNs are inherently ‘connected’
technologies; they require a stable, external endpoint to function. This directly
contradicts the goal of the Critical Infrastructure Fortify initiative, which is
to be able to isolate vital systems from harm. When we use a VPN to manage a **Programmable
Logic Controller** (PLC), we are not just providing access; we are exposing a pathway
that can be exploited to brick our infrastructure. Removing the VPN is an act of
fortification that **[prevents the total evaporation of security](https://www.agilicus.com/pwned-by-progress-the-year-your-air-gap-finally-evaporates/)**
when the air gap fails.

The solution to this architectural dilemma lies in becoming **disconnected yet connected**
through **Zero Trust Architecture**. At Agilicus, we believe that identity is the
new air gap. With **Agilicus AnyX**, we enable a model where access is granted not
to a network, but to a specific resource, and only after a rigorous, identity-based
verification process. This allows a water utility or power plant to remain isolated
from the public internet while still providing secure, granular access to authorized
personnel as needed. Because AnyX is clientless and operates at the application 
layer, it provides a level of utility and ease of use that traditional methods cannot
match. A technician can securely access a **Human-Machine Interface** (HMI) from
a standard browser without the friction of specialized software or hardware tokens.
Every connection is protected by **Multi-Factor Authentication** and recorded with
a complete audit trail, providing the visibility required for compliance with standards
like **[NERC CIP-003-9](https://www.agilicus.com/broken-barriers-why-johnny-cant-authenticate-and-why-nerc-will-fine-you/)**.
This architecture ensures that even if a business network is compromised, the vital
Operational Technology remains protected and reachable only by those whose identity
has been verified.

The Critical Infrastructure Fortify initiative is more than a set of guidelines;
it is a wake-up call for every operator of vital services. We can no longer afford
the luxury of assuming our perimeters are secure or our connections are stable. 
The path to resilience requires us to embrace isolation as a proactive strategy 
and to replace outdated, network-centric tools with identity-centric solutions. 
By shifting to a **Zero Trust** model, we increase the security of our most critical
assets while simultaneously improving the utility and efficiency of our operations.
The air gap may be evaporating, but by treating identity as our primary barrier,
we can build a fortress that is truly resilient against the threats of today and
tomorrow. We must move beyond the defensible perimeter and toward an architecture
that is defined by its ability to endure, recover, and continue delivering the essential
services our society relies upon.

## A Pragmatic Blueprint for Industrial Cyber Security

- Link: https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/
- Published: 2026-05-08T13:30:00+00:00
- Author: Don Bowman

![A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/
www/6d5ad600-blog-scorecard-7-wrapup-e15507f4-d4ad-423a-8d0f-300b0b0ecc9e.jpg)

Industrial Cybersecurity Best Practices

 * [Assessing Your Industrial Cyber Security Posture](https://www.agilicus.com/assessing-your-industrial-cyber-security-posture/)
 * [Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/)
 * [Identity and Credentials: The New Air Gap](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/)
 * [Halting Lateral Movement in Operational Technology](https://www.agilicus.com/halting-lateral-movement-in-operational-technology)
 * [System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure)
 * [Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network)
 * [A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/)

[Download the Best Practices Guide](https://www.agilicus.com/white-papers/industrial-cyber-security-best-practices/)

[Take the Cybersecurity Assessment](https://www.agilicus.com/l/industrial-cyber-security-best-practices-scorecard/)

Welcome to the final post in our seven-part series on industrial cyber security 
best practices. Over the last few weeks, we have unpacked the reality of securing
hybrid Information Technology and Operational Technology environments. We have moved
past the myth of the air gap and confronted the technical friction caused by connected
industrial control systems.

The core philosophy we have explored is the necessity of orthogonal defences. No
single security control is absolute. Every perimeter, no matter how robust, will
eventually face a threat it was not designed to handle. If you rely solely on a 
firewall, a stolen password renders it useless. If you rely solely on multi-factor
authentication, an unpatched vulnerability bypasses the login screen entirely.

To build true resilience, we must stack multiple, independent layers of defence.
As we have discussed, this is the Swiss cheese model in action. When the holes in
one layer are bypassed, the next layer must catch the threat using an entirely different
mechanism. Our program distils this strategy into five distinct dimensions.

### Reviewing the Five Dimensions

 1. **Boundary Defence**: We established that flat networks are an open invitation 
    for total compromise. By enforcing strict logical separation, eliminating open 
    inbound ports, and replacing legacy remote access with outbound-only reverse tunnels,
    we create structural friction at the edge of the network.
 2. **Identity and Credentials**: We recognised that identity is the new air gap. By
    mandating phishing-resistant multi-factor authentication, unifying corporate directories,
    and enforcing least privilege authorisation, we ensure that an attacker who bypasses
    the network boundary is immediately halted by robust identity controls.
 3. **Lateral Movement:** We addressed the scenario where a threat actor breaches the
    perimeter and steals a credential. By deploying network micro-segmentation and 
    eradicating legacy protocols like Remote Desktop Protocol, we contain the blast
    radius, preventing the attacker from navigating to critical process controllers.
 4. **System Hardening:** We looked at the assets themselves. By eradicating default
    credentials, disabling unused services, and implementing application allowlisting,
    we reduce the attack surface of individual devices, ensuring they are inherently
    resilient against exploitation.
 5. **Visibility and Detection:** Finally, we acknowledged that preventative controls
    will eventually fail. By implementing continuous asset discovery, industrial-aware
    network monitoring, and centralised log aggregation, we illuminate the network,
    allowing for rapid detection and response before a cyber event becomes a physical
    disaster.

### The Path Forward

These five dimensions are not a binary checklist to complete once and forget. They
represent a strategic architectural roadmap to govern your ongoing security investments.
They provide a pragmatic approach for mid-market companies to reduce risk and increase
control without crippling operational efficiency.

The threats facing critical infrastructure are not theoretical. As highlighted by
the Cybersecurity and Infrastructure Security Agency’s Risk and Vulnerability Assessments,
adversaries are actively exploiting weak boundaries and compromised credentials.
However, the defences are entirely within your control.

Your next step is to measure your current posture against this blueprint. We urge
you to take our [Cyber Security Assessment](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/#).
This scorecard will benchmark your environment against these five orthogonal dimensions,
highlighting your specific gaps. Once completed, you will be directed to download
the full Industrial Cyber Security Best Practices guide, providing the actionable
roadmap to address those vulnerabilities.

Do not wait for an incident to test your defences. Implement orthogonal controls,
secure your hybrid environment, and ensure the resilience of your operations. Take
the assessment today.

Industrial Cybersecurity Best Practices

 * [Assessing Your Industrial Cyber Security Posture](https://www.agilicus.com/assessing-your-industrial-cyber-security-posture/)
 * [Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/)
 * [Identity and Credentials: The New Air Gap](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/)
 * [Halting Lateral Movement in Operational Technology](https://www.agilicus.com/halting-lateral-movement-in-operational-technology)
 * [System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure)
 * [Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network)
 * [A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/)

[Download the Best Practices Guide](https://www.agilicus.com/white-papers/industrial-cyber-security-best-practices/)

[Take the Cybersecurity Assessment](https://www.agilicus.com/l/industrial-cyber-security-best-practices-scorecard/)

## Visibility and Detection: Illuminating the Industrial Network

- Link: https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network/
- Published: 2026-05-06T13:30:00+00:00
- Author: Don Bowman

![Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.
com/www/ea85520a-blog-scorecard-6-visibility-260bc02e-4192-4c0f-acec-479e3c00f8df.
jpg)

Industrial Cybersecurity Best Practices

 * [Assessing Your Industrial Cyber Security Posture](https://www.agilicus.com/assessing-your-industrial-cyber-security-posture/)
 * [Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/)
 * [Identity and Credentials: The New Air Gap](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/)
 * [Halting Lateral Movement in Operational Technology](https://www.agilicus.com/halting-lateral-movement-in-operational-technology)
 * [System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure)
 * [Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network)
 * [A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/)

[Download the Best Practices Guide](https://www.agilicus.com/white-papers/industrial-cyber-security-best-practices/)

[Take the Cybersecurity Assessment](https://www.agilicus.com/l/industrial-cyber-security-best-practices-scorecard/)

Welcome to the sixth post in our series on industrial cyber security best practices.
We have covered the preventative controls: boundaries, identity, lateral movement,
and system hardening. Today, we address the reality that preventative controls can,
and eventually will, fail. This brings us to the fifth orthogonal dimension: Visibility
and Detection.

You cannot protect what you cannot see. In many operational technology environments,
the network is a black box. If an adversary bypasses the perimeter and begins communicating
with a programmable logic controller using native industrial protocols, the business
is completely blind. Visibility and detection represent the crucial ability to identify
a breach in its early stages, allowing for rapid response before a cyber event becomes
a physical disaster.

### Asset Inventory and Discovery

The foundation of visibility is an accurate asset inventory. You cannot detect an
anomalous device if you do not know what devices are supposed to be there. In industrial
environments, rogue laptops, undocumented cellular modems, and forgotten test equipment
frequently blur the network topology.

Continuous, passive asset discovery is required to illuminate the environment. By
listening to network traffic, you can automatically build a dynamic inventory of
every device, its firmware version, and its communication patterns. This is the 
baseline upon which all detection capabilities are built.

### Network Security Monitoring

Traditional Information Technology monitoring focuses on corporate web traffic and
email. Operational Technology monitoring must understand industrial protocols like
Modbus, DNP3, and Ethernet/IP. An attacker altering the setpoint of a safety valve
will not trigger a standard corporate intrusion detection system; it simply looks
like standard network traffic.

Deploying network security monitoring specifically designed for industrial environments
is essential. These systems parse deep packet inspection to identify malicious commands,
firmware uploads, or abnormal reads and writes to process controllers. This deep
visibility is the only way to detect threat actors living off the land within the
industrial control system.

### Centralised Log Aggregation

When investigating an incident, piecing together logs from disparate firewalls, 
Windows servers, and identity providers is a painstakingly slow process. In a crisis,
time is the enemy. If logs are stored locally on individual machines, an attacker
can simply delete them to cover their tracks.

Centralised log aggregation solves this by securely forwarding all security events
to a unified platform. This allows security analysts to correlate events across 
the five dimensions. For example, a failed login attempt on the perimeter followed
by a configuration change on an internal workstation provides a clear picture of
an unfolding attack.

### Anomaly Detection and Alerting

Alert fatigue is a massive issue in cyber security. If a system generates thousands
of low-level warnings a day, the critical alerts will be ignored. Detection mechanisms
must be tuned to identify genuine anomalies, not just routine operational noise.

By establishing a baseline of normal communication patterns, anomaly detection engines
can highlight deviations. If an engineering workstation that typically communicates
with three specific programmable logic controllers suddenly attempts to scan the
entire subnet, this is a high-fidelity alert indicative of lateral movement.

### Incident Response Integration

Detection is useless without a coordinated response. When an alert fires indicating
a potential compromise of a safety instrumented system, the operations team and 
the security team must know exactly what to do.

Visibility mechanisms must integrate directly into the incident response plan. This
means defining clear escalation paths, establishing out-of-band communication channels,
and practicing tabletop exercises. The goal of visibility is to compress the mean-
time-to-repair, allowing operators to isolate the threat and restore normal operations
swiftly.

Visibility ensures that when the first four layers of defence are tested, you are
not caught off guard. Take our Cyber Security Assessment to evaluate your detection
capabilities, and download the best practices guide. In our final post, we will 
wrap up the series and bring all five dimensions together.

Industrial Cybersecurity Best Practices

 * [Assessing Your Industrial Cyber Security Posture](https://www.agilicus.com/assessing-your-industrial-cyber-security-posture/)
 * [Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/)
 * [Identity and Credentials: The New Air Gap](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/)
 * [Halting Lateral Movement in Operational Technology](https://www.agilicus.com/halting-lateral-movement-in-operational-technology)
 * [System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure)
 * [Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network)
 * [A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/)

[Download the Best Practices Guide](https://www.agilicus.com/white-papers/industrial-cyber-security-best-practices/)

[Take the Cybersecurity Assessment](https://www.agilicus.com/l/industrial-cyber-security-best-practices-scorecard/)

## System Hardening: Fortifying Industrial Infrastructure

- Link: https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure/
- Published: 2026-05-04T13:30:00+00:00
- Author: Don Bowman

![System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/
www/46036da4-blog-scorecard-5-hardening-3534dd44-06f2-41cc-8b56-639d68401565.jpg)

Industrial Cybersecurity Best Practices

 * [Assessing Your Industrial Cyber Security Posture](https://www.agilicus.com/assessing-your-industrial-cyber-security-posture/)
 * [Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/)
 * [Identity and Credentials: The New Air Gap](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/)
 * [Halting Lateral Movement in Operational Technology](https://www.agilicus.com/halting-lateral-movement-in-operational-technology)
 * [System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure)
 * [Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network)
 * [A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/)

[Download the Best Practices Guide](https://www.agilicus.com/white-papers/industrial-cyber-security-best-practices/)

[Take the Cybersecurity Assessment](https://www.agilicus.com/l/industrial-cyber-security-best-practices-scorecard/)

Welcome to the fifth post in our series exploring the dimensions of industrial cyber
security. We have built strong boundaries, secured identities, and restricted lateral
movement. Now, we must look at the assets themselves. This is the fourth orthogonal
dimension: System Hardening.

In operational technology, patching is notoriously difficult. Systems run constantly,
maintenance windows are rare, and legacy software often relies on outdated operating
systems. However, leaving systems in their default, vulnerable state is a massive
risk. System hardening is the process of reducing the attack surface of individual
devices, ensuring that even if an attacker reaches them, exploiting them is exceptionally
difficult.

### Eradicating Default Credentials

It is a harsh reality that many industrial control devices ship with hardcoded, 
publicly documented default passwords. Worse, these devices are often commissioned
and placed into production without these credentials ever being changed. An attacker
scanning the network only needs to consult a vendor manual to gain administrative
access to a critical process controller.

Eradicating default credentials is a non-negotiable first step in system hardening.
Every device must be configured with a unique, complex password before it is connected
to the operational network. This simple act removes the lowest-hanging fruit for
threat actors and forces them to use more complex, detectable methods.

### Disabling Unused Services

Every running service, open port, and enabled feature is a potential vulnerability.
Industrial devices often come with secondary web interfaces, file transfer protocols,
and legacy management services enabled by default, even if the operator never uses
them.

System hardening requires a ruthless audit of device configurations. If a service
is not strictly required for the operational function of the asset, it must be disabled.
By aggressively trimming the attack surface, you reduce the number of avenues an
attacker can exploit, significantly increasing the resilience of the individual 
component.

### Application Allowlisting

Traditional antivirus software relies on identifying known bad files, a strategy
that is increasingly ineffective against novel malware and custom ransomware. In
a static operational technology environment, where servers and human-machine interfaces
run the exact same software every day, we must invert this paradigm.

Application allowlisting ensures that only explicitly approved executables are permitted
to run. If an attacker manages to drop a malicious payload onto a hardened system,
the operating system simply refuses to execute it. This is a highly effective control
in industrial environments, locking down systems and preventing unauthorised code
execution.

### Secure Configuration Management

Configuration drift is a silent killer in industrial security. Over time, temporary
firewall rules become permanent, test accounts are forgotten, and security settings
are inadvertently downgraded during troubleshooting. Without a baseline, you cannot
know if a system is hardened.

Implementing secure configuration management involves defining a strict security
baseline for all asset classes and continuously monitoring against it. When a device
deviates from the baseline, it must trigger an immediate alert. This ensures that
the hardened posture you establish on day one is maintained throughout the lifecycle
of the equipment.

### Patch Management Strategy

While patching programmable logic controllers may require scheduled downtime, the
Windows-based servers and engineering workstations that manage them are frequent
targets and must be kept up to date. Relying on an air gap as an excuse not to patch
is a recipe for disaster.

A pragmatic patch management strategy acknowledges the operational constraints while
aggressively addressing critical vulnerabilities. This means triaging patches based
on actual risk, testing them in a representative environment, and deploying them
systematically. Where patching is impossible, compensating controls like micro-segmentation
must be enforced.

System hardening ensures that your foundational assets are resilient. To see where
your vulnerabilities lie, complete our Cyber Security Assessment and review our 
Industrial Cyber Security Best Practices guide. In our next post, we will explore
the final dimension: Visibility and Detection.

Industrial Cybersecurity Best Practices

 * [Assessing Your Industrial Cyber Security Posture](https://www.agilicus.com/assessing-your-industrial-cyber-security-posture/)
 * [Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/)
 * [Identity and Credentials: The New Air Gap](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/)
 * [Halting Lateral Movement in Operational Technology](https://www.agilicus.com/halting-lateral-movement-in-operational-technology)
 * [System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure)
 * [Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network)
 * [A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/)

[Download the Best Practices Guide](https://www.agilicus.com/white-papers/industrial-cyber-security-best-practices/)

[Take the Cybersecurity Assessment](https://www.agilicus.com/l/industrial-cyber-security-best-practices-scorecard/)

## Halting Lateral Movement in Operational Technology

- Link: https://www.agilicus.com/halting-lateral-movement-in-operational-technology/
- Published: 2026-05-01T13:30:00+00:00
- Author: Don Bowman

![Halting Lateral Movement in Operational Technology](https://www.agilicus.com/www/
a8a30ad0-blog-scorecard-4-lateral-2d3f8487-1c1b-48a8-9672-ba3fd943d62c.jpg)

Industrial Cybersecurity Best Practices

 * [Assessing Your Industrial Cyber Security Posture](https://www.agilicus.com/assessing-your-industrial-cyber-security-posture/)
 * [Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/)
 * [Identity and Credentials: The New Air Gap](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/)
 * [Halting Lateral Movement in Operational Technology](https://www.agilicus.com/halting-lateral-movement-in-operational-technology)
 * [System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure)
 * [Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network)
 * [A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/)

[Download the Best Practices Guide](https://www.agilicus.com/white-papers/industrial-cyber-security-best-practices/)

[Take the Cybersecurity Assessment](https://www.agilicus.com/l/industrial-cyber-security-best-practices-scorecard/)

Welcome to the fourth post in our series on industrial cyber security best practices.
We have covered [Boundary Defence](https://www.agilicus.com/halting-lateral-movement-in-operational-technology/#)
and [Identity and Credentials](https://www.agilicus.com/halting-lateral-movement-in-operational-technology/#).
Today, we examine the third orthogonal dimension: Lateral Movement.

If an attacker breaches the perimeter and steals a credential, their immediate next
step is to navigate the internal network. Lateral movement is the process by which
an adversary expands their foothold, moving from a compromised, low-value asset 
to critical control systems. This dimension addresses how easily a threat actor 
can move once inside. Defence in depth here means ensuring that even authenticated
sessions are contained, preventing broad network discovery and system-wide compromise.

### Network Micro-Segmentation

A flat internal network operates under the flawed assumption that everything inside
the perimeter is trusted. Once an adversary bypasses the external firewall, this
lack of internal friction allows them to move laterally with absolute impunity. 
They can discover and compromise critical assets without raising alarms.

Micro-segmentation divides the operational technology environment into tightly controlled
functional zones. By enforcing strict rules that govern which devices can communicate
with one another, you ensure that a compromised engineering workstation cannot freely
connect to unrelated process controllers. This approach contains the blast radius
of a breach. Divide the network by functional process areas and implement host-based
firewalls on all engineering workstations.

### Legacy Protocol Eradication

Legacy administrative protocols like the Remote Desktop Protocol and Windows Admin
Shares were designed for convenience, not security. Today, they are the primary 
mechanisms attackers use to navigate internal networks. Leaving these protocols 
enabled by default provides built-in highways for adversaries to deploy ransomware
across the entire industrial control system.

Eradicating or heavily brokering these protocols removes the native tools attackers
rely on to live off the land. This significantly elevates the difficulty of lateral
movement, forcing the attacker to generate noisy, detectable traffic. Disable Remote
Desktop Protocol internally unless explicitly required and brokered, and block the
use of cleartext protocols like Telnet.

### Application-Layer Brokering

When users authenticate to a traditional gateway, they are typically granted access
to a full subnet. This allows their machine to ping and map the surrounding network
architecture. This broad visibility is an intelligence goldmine for threat actors
conducting reconnaissance.

Application-layer brokering fundamentally alters this dynamic. Connections are established
only to the specific service requested, never dropping the user directly onto the
network itself. Because the network topology remains hidden from the endpoint, an
attacker who compromises a remote machine cannot scan the operational technology
environment. The network simply does not exist from their perspective.

### Least Privilege Routing

In many environments, routing tables are overly permissive, allowing any device 
to communicate with any other device across different subnets. This means a compromised
printer could theoretically initiate a connection to a human-machine interface. 
This is an unacceptable architectural flaw.

Implementing least privilege routing ensures that network paths only exist where
there is a documented operational necessity. By configuring switches and routers
to drop all traffic that does not match an explicit allow rule, you severely restrict
the pathways an attacker can use to traverse the environment.

### Securing Engineering Workstations

Engineering workstations hold the keys to the kingdom. They run the software used
to program and configure the programmable logic controllers that dictate physical
processes. If an attacker compromises one of these workstations, they effectively
own the plant floor.

These workstations must be treated as highly privileged assets, isolated from generic
corporate email and web browsing. By restricting lateral movement to and from these
critical machines, you ensure that a phishing email opened on a corporate laptop
cannot be used as a stepping stone to alter industrial processes.

Halting lateral movement ensures that a minor breach remains a minor breach. We 
encourage you to evaluate your internal network controls by taking our Cyber Security
Assessment and downloading the companion best practices guide. Next up, we will 
discuss the fourth dimension: System Hardening.

Industrial Cybersecurity Best Practices

 * [Assessing Your Industrial Cyber Security Posture](https://www.agilicus.com/assessing-your-industrial-cyber-security-posture/)
 * [Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/)
 * [Identity and Credentials: The New Air Gap](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/)
 * [Halting Lateral Movement in Operational Technology](https://www.agilicus.com/halting-lateral-movement-in-operational-technology)
 * [System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure)
 * [Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network)
 * [A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/)

[Download the Best Practices Guide](https://www.agilicus.com/white-papers/industrial-cyber-security-best-practices/)

[Take the Cybersecurity Assessment](https://www.agilicus.com/l/industrial-cyber-security-best-practices-scorecard/)

## Identity and Credentials: The New Air Gap

- Link: https://www.agilicus.com/identity-and-credentials-the-new-air-gap/
- Published: 2026-04-29T13:37:31+00:00
- Author: Don Bowman

![Identity and Credentials: The New Air Gap](https://www.agilicus.com/www/bbba571e-
blog-scorecard-3-identity-053498bd-8a4f-4446-b09f-f9be3ac1a828.jpg)

Industrial Cybersecurity Best Practices

 * [Assessing Your Industrial Cyber Security Posture](https://www.agilicus.com/assessing-your-industrial-cyber-security-posture/)
 * [Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/)
 * [Identity and Credentials: The New Air Gap](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/)
 * [Halting Lateral Movement in Operational Technology](https://www.agilicus.com/halting-lateral-movement-in-operational-technology)
 * [System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure)
 * [Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network)
 * [A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/)

[Download the Best Practices Guide](https://www.agilicus.com/white-papers/industrial-cyber-security-best-practices/)

[Take the Cybersecurity Assessment](https://www.agilicus.com/l/industrial-cyber-security-best-practices-scorecard/)

Welcome to the third post in our series on industrial cyber security best practices.
After establishing robust [Boundary Defence](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/#)
in our previous post, we now move to the second orthogonal dimension: Identity and
Credentials.

In an environment where network boundaries are increasingly porous, identity is 
the new air gap. When physical separation is no longer possible due to the convergence
of Information Technology and Operational Technology, robust authentication becomes
your primary line of defence. This dimension mandates a shift to an architecture
where identity is unified, heavily verified, and tied strictly to least privilege.

### Phishing-Resistant Multi-Factor Authentication

Relying on passwords alone is a proven failure. Legacy authentication methods, such
as short message service codes, are easily intercepted by modern threat actors through
techniques like sim-jacking. When administrative access to an infinitely important
role relies on weak authentication, the entire perimeter is compromised the moment
a single credential is stolen.

We believe phishing-resistant multi-factor authentication, such as hardware security
keys, is an absolute necessity. As we outlined in our article on [uncorrelated risks in multi-factor authentication](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/#),
tying a digital identity to a physical token ensures that stolen passwords cannot
be leveraged by a remote attacker. This adds a crucial layer of defence in depth.
Mandate hardware security keys for highly privileged access and disable email-based
authentication methods.

### Unified Identity and Single Sign-On

When operators manage hundreds of separate local accounts across disparate human-
machine interfaces, password reuse and credential sprawl become inevitable. This
operational friction results in dormant accounts remaining active long after an 
employee departs, providing a silent vector for exploitation.

Unifying identity through a single corporate directory eliminates this risk. Centralised
authentication allows for immediate, global revocation of access when personnel 
change, ensuring that the operational technology environment remains strictly governed
by a single source of truth. Disable local accounts on individual programmable logic
controllers and centralise all human identities.

### Third-Party and Vendor Access

Creating static, local accounts for third-party vendors typically results in shared
passwords and non-expiring access. This makes it impossible to confidently verify
who is actually connecting to your systems. As supply chains become more integrated,
managing the lifecycle of these external identities creates immense administrative
overhead.

Permitting vendors to authenticate using their own native corporate identity provider
shifts the burden of credential management back to them. You no longer store their
passwords; you simply authorise their verified identity to access specific resources
for a limited time. Eliminate shared generic accounts and implement time-bound access
that automatically expires.

### Granular Role and Resource Authorisation

Granting broad administrative access simply because an engineer needs to monitor
a single process violates the core tenet of least privilege. When users possess 
more access than required, any compromised account becomes a highly destructive 
tool in the hands of an attacker.

Granular authorisation ensures that roles are mapped tightly to specific tasks and
individual devices. By decoupling network access from application access, you ensure
that a contractor viewing read-only telemetry cannot inadvertently or maliciously
issue a stop command to a programmable logic controller. Restrict access on a strict
least-privilege basis.

### Machine-to-Machine Authentication

Static application programming interface keys are often embedded directly into scripts
and rarely rotated. They act as immortal passwords for automated systems. If an 
attacker extracts one of these keys from a compromised server, they gain unfettered,
persistent access to internal interfaces.

The modern approach requires abandoning static secrets in favour of short-lived 
tokens and continuous authorisation. By cryptographically binding machine identities
to specific hardware and requiring frequent token renewals, the window of opportunity
for an attacker to misuse a stolen credential is fundamentally eliminated.

Identity is not just about logging in; it is about verifying context continuously.
To see how your identity controls measure up, take our Cyber Security Assessment
and download the companion best practices guide. In our next instalment, we will
tackle the third dimension: Lateral Movement.

Industrial Cybersecurity Best Practices

 * [Assessing Your Industrial Cyber Security Posture](https://www.agilicus.com/assessing-your-industrial-cyber-security-posture/)
 * [Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/)
 * [Identity and Credentials: The New Air Gap](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/)
 * [Halting Lateral Movement in Operational Technology](https://www.agilicus.com/halting-lateral-movement-in-operational-technology)
 * [System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure)
 * [Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network)
 * [A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/)

[Download the Best Practices Guide](https://www.agilicus.com/white-papers/industrial-cyber-security-best-practices/)

[Take the Cybersecurity Assessment](https://www.agilicus.com/l/industrial-cyber-security-best-practices-scorecard/)

## Boundary Defence: The First Layer of Industrial Cyber Security

- Link: https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/
- Published: 2026-04-27T13:30:00+00:00
- Author: Don Bowman

![Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.
com/www/fe1202e5-blog-scorecard-2-boundary-c3dd3c40-f555-4673-bee8-bfda43f22d1c.
jpg)

Industrial Cybersecurity Best Practices

 * [Assessing Your Industrial Cyber Security Posture](https://www.agilicus.com/assessing-your-industrial-cyber-security-posture/)
 * [Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/)
 * [Identity and Credentials: The New Air Gap](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/)
 * [Halting Lateral Movement in Operational Technology](https://www.agilicus.com/halting-lateral-movement-in-operational-technology)
 * [System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure)
 * [Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network)
 * [A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/)

[Download the Best Practices Guide](https://www.agilicus.com/white-papers/industrial-cyber-security-best-practices/)

[Take the Cybersecurity Assessment](https://www.agilicus.com/l/industrial-cyber-security-best-practices-scorecard/)

Welcome to the second post in our seven-part series on industrial cyber security
best practices. In our [previous post](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/#),
we introduced the five orthogonal dimensions of our security scorecard, designed
for mid-market companies managing hybrid Information Technology and Operational 
Technology environments. Today, we focus on the first of these dimensions: Boundary
Defence.

Traditional perimeter security relied on the illusion of an air gap. The theory 
was simple: if the industrial control system is not physically connected to the 
outside world, it cannot be hacked. This was a comforting thought, but in modern
environments, the air gap is a dangerous myth. Economic pressures and operational
requirements demand remote monitoring, predictive maintenance, and vendor support.
The networks are connected.

This operational shift creates technical friction. When corporate networks bleed
into industrial environments, a single compromised workstation on the business side
easily cascades into a physical disruption on the plant floor. We must construct
hard boundaries.

### Multi-Layer Boundary Enforcement

The proliferation of interconnected devices means that a flat network is an open
invitation for total compromise. We must define strict physical and logical separation
between the corporate network and the operational technology network. This structural
separation forms the bedrock of defence in depth.

An attacker who breaches the corporate perimeter must face an entirely isolated,
heavily inspected environment before they can touch a programmable logic controller
or a safety instrumented system. This requires mandating firewall inspection at 
every boundary crossing, isolating safety systems entirely, and denying all direct
routing between the internet and operational technology assets.

### Elimination of Inbound Ports

Publicly exposed inbound ports are a liability. Search engines and automated scanners
actively exploit them. Relying on an open listener on your perimeter firewall guarantees
that adversaries will repeatedly test your defences until they find a weakness.

The architectural shift requires eliminating inbound ports completely. By forcing
all external integrations to rely on polled, outbound-only connections, you render
the operational technology environment invisible from the public internet. This 
removes the attack surface at the edge, drastically reducing the friction of constant
patching and external monitoring. Close direct inbound access for legacy remote 
desktop tools and replace them with outbound-only reverse tunnels.

### Modern Remote Access Architecture

Granting full network routing via legacy virtual private networks is equivalent 
to handing a contractor the master keys to the building just so they can fix a single
sink. Traditional jump boxes and shared remote access tools fail to provide the 
granular control required in modern industrial environments.

Access must be brokered through an identity-aware proxy that limits the user strictly
to the application they need. Terminating the session at the application layer breaks
the direct routing path. This means an infected remote machine cannot spread malware
laterally into the operational technology network. Deprecate legacy virtual private
network appliances and require continuous authorisation for every remote session.

### Alternate Perimeter Control

Vendors frequently install rogue cellular modems or undocumented wireless access
points to bypass strict corporate firewalls, prioritising maintenance convenience
over security. These alternate perimeters act as hidden backdoors, entirely subverting
the primary boundary defences you have meticulously designed.

Identifying and eliminating these undocumented connections is critical. By enforcing
strict network access control and physically auditing the environment, you ensure
that all traffic is forced through designated, heavily inspected chokepoints. Conduct
physical audits and block unauthorised vendor backdoors installed during commissioning.

### Outbound Traffic Restriction

A compromised internal device must communicate with an external command and control
server to receive instructions or exfiltrate data. If outbound traffic from the 
operational technology network is permitted by default, attackers face zero resistance
once inside.

Implementing a strict default-deny posture for outbound traffic turns this vulnerability
into a containment trap. By explicitly defining which internal devices are allowed
to communicate externally and filtering that traffic through deep inspection, you
ensure that malicious beacons fail to connect, buying time for the security team
to respond.

Boundary defence is not about building an impenetrable wall; it is about creating
structural friction that isolates critical assets. It is the first layer of our 
orthogonal defence strategy. We encourage you to evaluate your own boundary defences
by taking our Cyber Security Assessment and downloading the Industrial Cyber Security
Best Practices guide. In our next post, we will explore the second dimension: Identity
and Credentials.

Industrial Cybersecurity Best Practices

 * [Assessing Your Industrial Cyber Security Posture](https://www.agilicus.com/assessing-your-industrial-cyber-security-posture/)
 * [Boundary Defence: The First Layer of Industrial Cyber Security](https://www.agilicus.com/boundary-defence-the-first-layer-of-industrial-cyber-security/)
 * [Identity and Credentials: The New Air Gap](https://www.agilicus.com/identity-and-credentials-the-new-air-gap/)
 * [Halting Lateral Movement in Operational Technology](https://www.agilicus.com/halting-lateral-movement-in-operational-technology)
 * [System Hardening: Fortifying Industrial Infrastructure](https://www.agilicus.com/system-hardening-fortifying-industrial-infrastructure)
 * [Visibility and Detection: Illuminating the Industrial Network](https://www.agilicus.com/visibility-and-detection-illuminating-the-industrial-network)
 * [A Pragmatic Blueprint for Industrial Cyber Security](https://www.agilicus.com/a-pragmatic-blueprint-for-industrial-cyber-security/)

[Download the Best Practices Guide](https://www.agilicus.com/white-papers/industrial-cyber-security-best-practices/)

[Take the Cybersecurity Assessment](https://www.agilicus.com/l/industrial-cyber-security-best-practices-scorecard/)
