Topics Tools Books & Guides Talk to Securafy

Knowledge Hub / Cybersecurity

Cybersecurity

Your EDR Can Be Blinded: Why BYOVD Attacks Mean Endpoint Tools Aren't Enough

BYOVD attacks use signed but vulnerable drivers to kill EDR before ransomware runs. See which layers still protect SMBs when the endpoint agent goes dark.

Rodney Hall By Rodney Hall Updated Nov 2026 11 min read Share

If ransomware operators can switch off your endpoint protection before they encrypt anything, what exactly are you paying for? That question has stopped being hypothetical. BYOVD attacks, short for Bring Your Own Vulnerable Driver, let attackers load a legitimate but flawed Windows driver and use its kernel-level access to kill or blind EDR software. ESET Research now tracks almost 90 EDR killers actively used in the wild, 54 of them built on BYOVD and abusing 35 different vulnerable drivers.

The direct answer: EDR is still essential, but it cannot be your only source of visibility, because a BYOVD attack is designed to remove it. Businesses need layers that keep working when the endpoint agent goes quiet: tight admin privileges, driver and application control, alerts when an agent stops reporting, network-level telemetry, and a human team that responds within minutes.

How do BYOVD attacks blind EDR?

A BYOVD attack works by abusing trust in signed drivers. The attacker, who already has administrator rights, installs a legitimate driver that contains a known flaw. Because the driver is signed, Windows loads it into the kernel, and the attacker then uses the flaw to terminate security processes from a level of the operating system that the EDR agent cannot defend against.

ESET describes EDR killers as a predictable stage of modern ransomware intrusions: the affiliate gains high privileges, runs a tool to disrupt protection, and only then launches the encryptor. That sequencing is deliberate. Encryptors are noisy, so rather than making them stealthy, attackers clear the room first. ESET's guide for business leaders on EDR killers calls these tools cheap, accessible, and increasingly plug-and-play.

Is this a niche technique or a mainstream one?

It is mainstream and commercialized. SentinelOne's research on how EDR killing became an industry traces a shift from one-off tools to maintained, versioned products shared across ransomware crews, with a single tool reportedly used in Black Basta, LockBit, BlackCat, and other operations. In June 2026, ESET published an investigation into the Gentlemen ransomware group's EDR killer framework, finding that the gang's operators build and supply an EDR killer suite to affiliates, with an in-house tool that targets more than 400 processes across 48 security products.

Bitdefender's 2026 analysis of ransomware targeting US organizations adds a timing concern: EDR-blinding is increasingly delivered in a single stage, with some ransomware now shipping with a vulnerable driver already embedded. When defense evasion and encryption happen almost together, the window for a human to intervene shrinks.

A real intrusion: an expired forensic driver and a VPN account without MFA

The clearest recent example comes from Huntress. In early February 2026, Huntress investigated an intrusion in which the attacker logged in with compromised SonicWall SSL VPN credentials on an account that lacked MFA. After reconnaissance, the attacker deployed an EDR killer disguised as a firmware update utility. It dropped a legitimate EnCase forensic driver whose signing certificate had expired in 2010 and been revoked, which Windows still accepted because of how older signed drivers are validated.

The tool carried a list of 59 security process names and checked for them every second, killing any it found. Huntress caught and contained the intrusion before ransomware was deployed, in part by using the SonicWall firewall telemetry alongside endpoint visibility. Two lessons stand out for SMB leaders. The attack started with a basic identity gap, not an exotic exploit. And it was stopped because detection did not depend on a single endpoint agent surviving.

The uncomfortable truth: more EDR is not the answer

When business owners ask whether they should switch to a "better" EDR after reading about BYOVD, our answer is usually no. The problem is architectural, not a matter of brand. Any agent running on a machine where the attacker has administrator rights and kernel access is fighting from a weaker position. Swapping one agent for another changes which process names the EDR killer targets, not whether it can try.

ESET makes a related point that challenges the most common defensive instinct. Blocking vulnerable drivers from loading is necessary, but it happens at the last possible moment: by the time the attacker tries to load the driver, they already hold high privileges and are seconds from launching the encryptor. If one EDR killer fails, they try another. ESET also documents driverless EDR killers that block an agent's communication with its cloud backend or freeze its processes, which no driver blocklist will stop.

So the useful question is not "which EDR can't be killed?" It is "what still sees the attacker, and who responds, after the EDR has been killed?"

What should SMBs do about BYOVD attacks?

The defensive plan has to move earlier in the attack and add visibility that does not live on the endpoint. Each layer below covers a different gap, and none of them is sufficient alone:

LayerWhat it stops or revealsWhat it misses
MFA on VPN and remote accessThe initial login, as in the Huntress caseAttackers who arrive through phishing or an exposed service
Least privilege for admin rightsThe privilege needed to install a kernel driverAttackers who steal an existing admin account
Driver blocklist and memory integrityKnown vulnerable drivers at load timeUnlisted drivers and driverless EDR killers
Default-deny application controlThe EDR killer executable before it runsAbuse of tools that are already approved
Agent health alerts and network telemetryAn EDR going silent and lateral movement on the wireNothing on its own; it needs a responder

On the driver layer specifically, Microsoft states that its vulnerable driver blocklist has been enabled by default since the Windows 11 2022 update and is also enforced when memory integrity (HVCI), Smart App Control, or S mode is active. It is updated quarterly and through monthly Windows updates. That means older Windows 10 machines, devices where memory integrity was switched off for compatibility, and systems that are behind on updates are exactly where the protection is weakest. SentinelOne notes that many drivers in the public LOLDrivers catalog are absent from the default blocklist, which is why application control and monitoring still matter.

A practical action list for SMB leaders and their IT providers:

  • Confirm MFA is enforced on every VPN, firewall, and remote access account, with no service or legacy exceptions.
  • Remove standing local and domain admin rights from day-to-day accounts so loading a driver requires a separate, monitored credential.
  • Verify memory integrity and the vulnerable driver blocklist are on across your Windows fleet, and get a list of any machines where they are off and why.
  • Ask how long it takes for someone to be alerted when an EDR agent stops checking in, and who acts on that alert outside business hours.
  • Make sure your incident response plan assumes the EDR may be disabled mid-attack and names the other data sources responders will use.

Why does response speed matter more than detection?

Because ransomware intrusions are human-operated, detection only helps if someone acts on it immediately. ESET's researchers stress that defenders, whether an internal SOC, an MSSP, or an MDR provider, have to respond "correctly, immediately, and with sufficient decisiveness," since attackers adapt when a step fails. An alert that an EDR agent went offline at 2 a.m. is worth very little if it sits in a queue until Monday.

This is where many SMB security stacks quietly fall short. They have an EDR license and perhaps an MDR add-on, but nobody has defined who isolates the machine, who disables the compromised account, and how quickly. We cover that ownership question in detail in what SMBs need after the MDR alert fires. The same logic applies if you run on Microsoft's built-in tooling, as explained in what Defender and Azure MFA do not solve alone.

The business stakes are straightforward. A disabled EDR followed by encryption means downtime, recovery cost, and possible data theft. Our breakdown of the real dollar cost of downtime and ransomware for SMBs shows why minutes of response time translate directly into days of disruption.

Does AI change the BYOVD picture?

It is starting to. ESET says it strongly suspects AI assisted in developing some EDR killers it analyzed, including a tool used by the Warlock group that cycles through device names in a trial-and-error pattern. Lower development effort means more variants and faster iteration, which reinforces the case for controls that block behavior and unapproved software rather than chasing individual samples. For broader context, see how cybercriminals use AI to launch smarter attacks and our AI Lab guide to AI security risks and controls for small businesses.

How Securafy approaches EDR killers

Our approach starts before the driver ever loads. Default-deny application control means an unrecognized executable, such as an EDR killer posing as a firmware utility, is blocked from running in the first place, and we separate administrative credentials from everyday accounts so a single compromised login does not grant the rights BYOVD requires. We also enforce MFA on remote access paths like VPNs, which would have closed the door in the Huntress case.

Behind that, our 24/7 security operations team treats an endpoint agent that stops reporting as an incident, not a maintenance ticket, and correlates firewall and identity signals so visibility survives if the agent does not. The goal is simple: if one layer is disabled, the next one is already watching and a human is already moving.

Where To Go From Here

BYOVD attacks turn endpoint protection into a target, so the measure of your security is what still works after the EDR is gone. Ask your IT provider how quickly they would know if your agent went silent and what they would do next.

To see how prevention-first controls, monitoring, and response work together, review Securafy's managed security services.

If you would rather talk through your specific environment first, book a strategy call with Securafy and we will review your endpoint, privilege, and remote access setup against the techniques ransomware crews are using now.

Tagged Under Cybersecurity

Join The Conversation

Have a question or perspective on this topic? Add it below.

Rodney Hall

About The Author

Rodney Hall · President & COO

Rodney Hall is the President and COO of Securafy, with 2 decades of experience in IT service management and operations.

He writes about the less glamorous but essential side of IT: support systems, documentation, business continuity, recurring issues, downtime, and the processes that keep client environments running well. His perspective comes from years spent improving how service is delivered, how teams respond, and how small problems are prevented from becoming much larger ones.

Outside of work, Rodney enjoys home improvement projects, woodworking, and dirt bike riding. His personal mission mirrors Securafy’s: helping businesses stay secure, compliant, and ready for whatever comes next.

Writes about: Managed IT, IT operations, service delivery, business continuity, downtime prevention, support processes, operational risk

More From This Author →

Get Practical Cybersecurity Field Notes

Monthly cybersecurity, compliance, and IT strategy updates from Securafy, written for business owners who need clear next steps.

  • Practical security tips from our Cyber Security Drip series
  • The Securafy Times, our monthly roundup on compliance and IT strategy
  • Occasional updates on new tools, guides, and research
  • No spam — unsubscribe anytime