Outlook under attack: Microsoft's November patch closes an exploited flaw that leaks NTLMv2 hashes

On 10 November 2026, Microsoft released fixes for more than a thousand vulnerabilities. One of them, CVE-2026-124194 in Outlook, was already being exploited by attackers before Microsoft had a patch ready. Opening a crafted message file can expose a user's password hash, which an attacker can then use to impersonate that user. Every device running Outlook needs the update.

Microsoft Security News 4 min read

Outlook under attack: Microsoft's November patch

On 10 November 2026, Microsoft's regular Patch Tuesday release fixed 1,094 CVEs across its products. One flaw in the list stands out. CVE-2026-124194 is a spoofing vulnerability in Microsoft Outlook, and Microsoft has marked it as actively exploited. The company confirms that attacks took place before the fix was released.

The flaw had not been publicly disclosed before the patch. By 11 November, security publications were already describing it as a zero-day. That term means a vulnerability that attackers use before any fix exists. It matters because, for this kind of flaw, the usual gap between a patch being published and being installed gives attackers an opening.

Item

Details

Identifier

CVE-2026-124194

Affected products

Microsoft Outlook in Office LTSC 2021, Office LTSC 2024 and Microsoft 365 Apps for Enterprise (32-bit and 64-bit)

Microsoft rating

Moderate, CVSS 6.5

Weakness type

CWE-356: the software fails to warn users before an unsafe action

Exploitation status

Exploited before the patch was released

CISA KEV catalogue

Added on 10 November 2026. Reported remediation deadline for US federal agencies: 24 November 2026

Are you affected?

Check whether your organisation runs any of the Office versions listed above. If it runs none of them, this flaw does not apply to you.

The attack needs two things. The attacker has to be able to send a message to the user, and the user has to open the crafted file in Outlook. Microsoft states that the Preview Pane is not an attack vector, so simply selecting a message in the list does not trigger the leak.

Do not rely on that alone. Opening attachments and message files is a routine task in many organisations, so "our users would never do that" is not a safe assumption.

Fix it

  1. Update Office, not only Windows. The fix for this flaw is not delivered through Windows Update. It comes through the Office update mechanism.

  2. On an individual computer, open Outlook or Word and go to File > Account > Update Options > Update Now. In a Polish-language interface the path is Plik > Konto > Opcje aktualizacji > Aktualizuj teraz. Many IT teams work with English interfaces, so both labels are given here.

  3. In a managed environment, force an update through Click-to-Run channels or Microsoft 365 Apps update policies. Confirm that all devices run the build listed in Microsoft's November 2026 security guidance.

  4. Block outbound SMB traffic. The attack relies on Outlook connecting to a server controlled by the attacker, typically over SMB on TCP port 445. Blocking outbound port 445 to external addresses reduces the risk of hash leakage while patching is still under way.

Has it already happened to you?

Microsoft confirms exploitation but does not say who was targeted or how many attacks took place. There is no public evidence that any particular organisation was compromised. It is still worth checking your logs.

Look for outbound connections from machines running Outlook to unfamiliar external hosts, particularly on port 445, in firewall and endpoint detection logs. Also look for unusual NTLM authentication attempts that do not match normal internal traffic.

Why it matters more than its score

A CVSS score of 6.5 looks moderate, and Microsoft has rated the flaw Moderate. A CVSS score, however, measures theoretical technical impact, not real-world risk. This flaw is already being exploited, and its result is the theft of login material. An NTLMv2 hash does not always give an attacker full control of an account, but it lets them attempt authentication against other systems, particularly where NTLM is still enabled.

The same release also contains other Outlook flaws rated Critical. None of them is known to be exploited. For an IT team, the order of work is therefore clear: patch the flaw under active attack first, then the rest, even though some of those ratings look more severe on paper.

The case also shows how much weight NTLM still carries in many networks. Wherever NTLM runs quietly in the background, any flaw that exposes its hashes becomes more dangerous. Reducing NTLM use where possible limits the damage from similar bugs in future.

Blog

Security 4 min read

Autopay confirms unauthorised access to payment infrastructure: what is known and what customers should check

Autopay, a Polish online payment operator, has confirmed that an unauthorised party gained access to part of its infrastructure and to some payment-related data. The company detected the intrusion on 11 November 2026 and announced it on 13 November. It has not said whether data was copied, how many customers are affected, or which data categories were involved. Customers should check their bank statements now, because the company itself cannot rule out data loss.

AI 6 min read

$740,000 on Someone Else's Keys

An attacker broke out of Railway's build system, reached the service that decrypts customer secrets, and searched the variable names for anything that looked like an AI provider key. The bill arrived at the customers. Railway has since said it will pay it.

AI 6 min read

OpenAI Rated It Critical, Then Shipped It

GPT-6 Cyber finds vulnerabilities and writes working exploits. OpenAI gave it the highest cyber risk rating its own framework allows, and released it anyway, to a vetted list. The argument for doing so rests on a number that has never been tested outside a benchmark.