Skip to main content

How to Check Whether a Windows 11 PC Has Actually Been Compromised

Media How to Check Whether a Windows 11 PC Has Actually Been Compromised

 

Your Windows 11 PC suddenly behaves strangely.

The fans run harder than usual. An unfamiliar process appears. Microsoft Defender raises an alert. Your browser redirects somewhere unexpected, or you receive a Microsoft 365 authentication request you never initiated.

Has the PC genuinely been compromised?

Or is Windows simply:

installing updates

indexing files

synchronising OneDrive

scanning with Defender

updating applications

performing legitimate background maintenance


There is no single button that can prove a Windows PC is completely safe.

A proper compromise assessment combines evidence from the device, user accounts, Microsoft 365, email, security tools and network.

And if an attack may still be active, the first priority is not scanning.

It is containment.

First: Don't Treat Every Slow PC as Hacked

High CPU usage or lots of Task Manager processes are not evidence of compromise by themselves.

Windows 11, Microsoft 365, browsers, OneDrive, security software and hardware-management tools can all run numerous background processes legitimately.

Concern should rise when several suspicious indicators occur together.

For example:

Slow PC alone: probably needs investigation.

Slow PC + Defender disabled + unknown administrator account + unfamiliar Microsoft 365 sign-ins: potentially serious security incident.

Record:

what happened

when it started

who was signed in

any Defender alerts

recent downloads or support calls

whether other users are affected

whether unusual authentication activity has appeared


A timeline becomes extremely useful if the problem turns out to be genuine.

Warning Signs That Deserve Immediate Attention

Take the situation more seriously if you see:

ransomware or credential-stealer detections

Defender disabled unexpectedly

unfamiliar antivirus exclusions

unknown administrator accounts

unexpected MFA requests

passwords changing unexpectedly

unfamiliar Microsoft 365 sign-ins

unapproved remote-access software

browser extensions returning after removal

PowerShell windows appearing without explanation

files suddenly changing extension

ransom notes

customers receiving messages you didn't send

unknown Outlook forwarding rules

new mailbox delegates or administrators

similar suspicious behaviour across multiple computers


One clue may have an innocent explanation.

Several together usually deserve proper incident investigation.

1. If the Attack May Be Active, Contain It First

If files are actively being encrypted, accounts appear to be under somebody else's control or the problem appears to be spreading:

Stop normal use of the PC.

For an unmanaged machine:

1. Disconnect Ethernet.


2. Turn off Wi-Fi.


3. Disconnect VPN.


4. Disconnect unnecessary external drives.


5. Do not attach backup storage.


6. Contact IT using another trusted device.

 

The NCSC recommends disconnecting infected devices from network connections to reduce the potential for further damage or spread.

For a business using Microsoft Defender for Endpoint, IT may instead isolate the device remotely. Microsoft says device isolation restricts network communication while maintaining connectivity to Defender for Endpoint so the security team can continue investigating it.

That can be preferable to an employee manually experimenting with the machine.

2. Preserve Evidence Before You Start “Cleaning”

This is a point I would strengthen considerably.

If the incident may be serious, don't immediately delete every suspicious file or reset Windows.

Useful evidence can include:

Defender detection names

timestamps

screenshots

suspicious email messages

unfamiliar applications

file paths

ransom notes

account sign-in records

IP addresses

browser history

relevant logs


Deleting everything first can make it much harder to answer:

How did the attacker get in?

What did they access?

Are other computers affected?

Did they steal credentials?

Could the same attack happen again?

For a serious business compromise, containment and evidence collection should normally come before aggressive cleanup.

3. Check Microsoft Defender

Open:

Windows Security > Virus & threat protection

Confirm which antivirus provider is active.

If Microsoft Defender is supposed to be protecting the computer, check that:

real-time protection is enabled

cloud-delivered protection is enabled

security intelligence is current

tamper protection is enabled

no unexpected warning is present


An unexplained security product being disabled is much more significant than a PC simply feeling slow.

Review Protection History

Open:

Windows Security > Virus & threat protection > Protection history

Look for:

detected malware

failed remediation

threats that were allowed

potentially unwanted applications

security controls being disabled

recurring detections


Record the:

detection name

file path

timestamp

action taken


before changing anything.

A detection marked Remediated does not necessarily mean the whole incident is over.

A credential-stealing Trojan could have captured passwords before Defender removed it.

4. Check for Suspicious Defender Exclusions

Open:

Virus & threat protection > Manage settings > Exclusions

Broad unexplained exclusions deserve investigation.

Examples might include:

C:\

C:\Users

C:\Windows

or whole temporary/script directories.

Legitimate business software can require antivirus exclusions, so don't automatically remove them on a managed PC.

But somebody should be able to explain why each exclusion exists.

Attackers sometimes weaken protection rather than disabling Defender completely.

5. Update Defender and Scan Appropriately

If the situation is not an active incident requiring forensic handling, update Defender and scan.

Microsoft recommends scanning when you suspect malware or want to confirm that a previous infection appears to have been removed.

Start with a Quick scan for ordinary suspicion.

Use a Full scan where stronger indicators exist.

For persistent threats, Microsoft Defender also provides Offline scan, which restarts the computer and scans outside the normal Windows session.

But remember:

A clean scan is evidence. It is not proof that compromise never occurred.

6. Check Recently Installed Applications

Go to:

Settings > Apps > Installed apps

Sort or review by installation date.

Look particularly for:

unfamiliar remote-support software

browser assistants

unknown VPN clients

screen-sharing software

cryptocurrency miners

questionable “driver updater” utilities

applications installed during a suspicious support call


Remote-access software isn't automatically malicious.

Tools legitimately used by IT support can also be abused by scammers.

The question is:

Who installed it, when and why?

7. Review Accounts and Administrators

Open:

Settings > Accounts > Other users

Investigate:

accounts nobody recognises

old support accounts

former employee accounts

standard users unexpectedly promoted to administrator

accounts created around the suspected compromise date


An unknown administrator account is a serious finding.

Do not simply delete it and assume the incident is over.

An attacker with administrator access may have made many other changes.

8. Check Browser Extensions and Fake Notifications

Not every scary popup is malware.

A website granted browser-notification permission can display convincing fake virus alerts even after you've left the site.

Review each browser's:

extensions

site notification permissions

search engine

homepage

startup pages

proxy configuration


Investigate extensions you don't recognise.

And never telephone a number displayed by a random browser warning claiming your computer is infected.

9. Check the User's Identity From Another Trusted Device

This is one of the most important additions to any endpoint investigation.

A PC can be completely clean while the user's account is already compromised.

For Microsoft personal accounts, review recent account activity.

For Microsoft 365 work accounts, review recent sign-ins and involve IT so Microsoft Entra logs can be checked more deeply.

Look for:

successful sign-ins you don't recognise

unfamiliar devices

unexpected MFA activity

authentication-method changes

password resets

suspicious geographic or network patterns


Don't rely solely on geography.

VPNs, corporate networks and mobile providers can make legitimate sign-ins look unusual.

Look at the combination of:

time + device + authentication method + location + user behaviour.

10. Check Microsoft 365 Mailbox Persistence

Changing a password may not remove everything an attacker configured.

Inspect:

Inbox rules

automatic forwarding

mailbox delegates

Sent Items

Deleted Items

connected applications

registered authentication methods


Attackers may create rules that:

forward financial conversations

hide security warnings

delete replies

monitor invoices

conceal password-reset messages


For businesses, Microsoft 365 audit information is much more useful than simply examining what appears in Outlook.

A compromised mailbox is an identity incident, not merely an email problem.

11. Check for Ransomware

Immediate ransomware indicators include:

large numbers of files changing rapidly

documents becoming unreadable

unfamiliar file extensions

ransom notes

network shares suddenly failing

backup errors

security products being disabled


If ransomware is present:

Contain first.

Do not connect backup storage.

Do not immediately start restoring files to systems that may still be compromised.

The NCSC advises recovering only onto environments you are confident are clean and verifying that backups are not infected before restoration.

12. A Clean Antivirus Scan Does Not Prove the PC Wasn't Compromised

This should be one of the article's central messages.

A clean scan can tell you:

Defender did not detect known active malware at that moment.

It cannot prove:

malware was never present

credentials weren't stolen

data wasn't copied

an attacker didn't use legitimate remote tools

Microsoft 365 wasn't accessed

malicious mailbox rules weren't created

persistence hasn't been missed

another PC isn't compromised


Modern attackers increasingly use legitimate credentials and legitimate administrative tools.

That can leave little conventional malware behind.

So the correct question isn't merely:

“Does antivirus find anything?”

It is:

“What evidence exists across endpoint, identity, email and network activity?”

13. When Should You Rebuild Windows?

Sometimes trying to clean the system is not worth the uncertainty.

A controlled rebuild becomes attractive when:

a backdoor or remote-access Trojan was present

an attacker obtained administrator/SYSTEM access

security controls were deliberately disabled

persistence mechanisms are found

ransomware affected the PC

malware repeatedly returns

system integrity cannot be established

the device handled particularly sensitive information

regulatory or contractual requirements demand a stronger assurance level


The NCSC says that if antivirus cannot reliably clean an infected device, wiping it and reinstalling the operating system may be necessary.

For ransomware recovery, the NCSC likewise recommends safely wiping infected devices and reinstalling the operating system before restoring clean data.

Don't Restore the Problem

Before rebuilding, preserve essential business data carefully.

But do not blindly copy the entire old user profile—including every script, extension and application-data directory—back onto the clean machine.

You may simply reintroduce the same unwanted configuration.

Restore required data deliberately.

What Not to Do

Avoid installing several antivirus products because you think “more scanners = more certainty”.

Avoid random malware-removal utilities.

Don't change important passwords on a computer you suspect contains a keylogger.

Don't attach backup drives to an actively compromised system.

And don't factory-reset the PC before considering whether the incident involves:

stolen credentials

Microsoft 365 sessions

mailbox rules

other devices

data theft


Resetting one laptop does not undo everything an attacker may already have done elsewhere.

A Better Compromise Checklist

For a suspected Windows 11 compromise, I would simplify the process to:

1. Decide whether the attack may still be active.


2. Contain the PC immediately if necessary.


3. Preserve useful evidence.


4. Check Defender health, history and exclusions.


5. Update security intelligence and run appropriate scans.


6. Review recently installed software and administrator accounts.


7. Inspect browsers for suspicious extensions and permissions.


8. From a trusted device, review account sign-ins and MFA.


9. Check Microsoft 365 mailbox rules, forwarding and delegated access.


10. Look for ransomware or unusual file activity.


11. Check whether other users or devices show similar indicators.


12. Decide whether the PC can genuinely be trusted again.


13. Rebuild it when integrity cannot be established.


14. Reset credentials and revoke sessions from a trusted environment.


15. Fix the route the attacker originally used.

 

That is easier to follow than treating compromise assessment as 24 independent checks.

Prevention After Recovery

Once the incident is contained, ask:

Why did it happen?

Potential routes include:

phishing

stolen passwords

weak MFA

unpatched software

malicious attachments

remote-access tools

exposed services

compromised suppliers

another infected device


Then strengthen the relevant controls.

For a business, that might include:

phishing-resistant MFA/passkeys

Microsoft Defender for Business

Defender for Endpoint

Conditional Access

Microsoft Intune

least privilege

application control

patch management

security monitoring

tested backups

staff awareness


Incident response should end with risk reduction, not simply the disappearance of a warning.

How Hamilton Group Can Help

Checking whether a Windows 11 PC has genuinely been compromised requires much more than running an antivirus scan.

Hamilton Group can investigate the endpoint alongside the user's Microsoft 365 identity, mailbox, security configuration and wider business environment.

We can help with:

rapid containment

Microsoft Defender investigations

endpoint detection and response

suspicious account activity

Microsoft Entra sign-ins

mailbox compromise

ransomware

credential theft

Windows rebuilds

root-cause analysis

long-term security improvements


Hamilton Group aims to make first contact on IT support requests within 15 minutes, so businesses can get experienced guidance quickly when something genuinely suspicious happens.

Visit hgmssp.com or call 0330 043 0069 to discuss a suspected compromise.