On this page
On 10 November 2026 Microsoft fixed a flaw in classic Outlook that attackers were already using. A user opens a crafted message file, Outlook quietly connects to a server the attacker controls, and Windows hands over the user's NTLM credential hash on the way.
If your organisation runs Microsoft 365 Apps or Office LTSC on Windows, this one is yours. The fix ships through Office updates, not Windows Update, so a fully patched Windows estate can still be exposed. This brief covers what is known, how to check your exposure, how to look for signs it was used against you, and what to do this week.
What happened
Microsoft's November 2026 security release, published on Patch Tuesday (10 November), includes CVE-2026-124194, described as a spoofing vulnerability in Microsoft Office Outlook. Microsoft rates it Moderate with a CVSS score of 6.5 and flags it as exploited in the wild. The advisory is marked "Customer Action Required".
The same day, CISA added the CVE to its Known Exploited Vulnerabilities catalogue. US federal agencies must remediate it by 24 November 2026. That deadline does not bind anyone else, but KEV listing is a strong signal: someone has seen it used in real attacks, not just in a lab.
As reported, the attack needs the victim to open a crafted message file. When they do, Outlook reaches out to a network location the attacker controls without showing the usual warning about external content. Windows then tries to authenticate to that location, and in doing so sends the user's NTLMv2 challenge-response.
Do not read the "Moderate" rating as "low priority". The score reflects that a user has to act. In practice, getting someone to open an attachment is the most reliable step in most phishing campaigns, and the payoff here is a reusable credential.
Who is affected
Based on what has been published so far:
- Affected: classic Outlook for Windows, across supported versions. Reporting names Microsoft 365 Apps for enterprise and Office LTSC among the affected channels.
- Fix delivery: Office updates (Click-to-Run channels or the relevant MSI update), not the Windows cumulative update. This is the detail most likely to leave gaps. Teams that track patch compliance through Windows Update reports will see green while Outlook stays vulnerable.
- Highest impact: environments where NTLM is still broadly accepted, SMB and LDAP signing are not enforced, and user devices can reach the internet on TCP 445. A cracked password also matters more where that password is reused on systems without multi-factor authentication.
We have not seen confirmed reports that the new Outlook for Windows, Outlook on the web or Outlook for Mac are affected. Check Microsoft's advisory for the authoritative product list rather than relying on this summary.
How to check your exposure
Work through these in order. The first two tell you whether you are vulnerable; the rest tell you how bad a leaked hash would be.
1. Which machines run classic Outlook, and on which build
In Outlook, File → Office Account → About Outlook shows the version and build. At scale, use your endpoint management tool (Intune, Configuration Manager or similar) to report installed Office build and update channel per device. Compare against the November 2026 fixed builds listed in Microsoft's release notes for each channel.
Pay attention to devices on slower channels. Semi-Annual Enterprise Channel and LTSC machines often lag weeks behind, by design.
2. Whether Office updates are actually applying
Look for devices where Office updates are disabled by policy, pointed at a stale internal source, or blocked because Outlook has not been restarted. A downloaded update that waits for the user to close Outlook is not a deployed fix.
3. Whether user devices can reach the internet on SMB
From a normal user workstation, outbound TCP 445 to the internet should fail. If it succeeds, any NTLM leak that targets an SMB path reaches the attacker directly. WebDAV over HTTP or HTTPS is a second route that some NTLM leaks use, so do not stop at port 445.
4. How much NTLM your network still accepts
If domain controllers and file servers require SMB signing, and LDAP signing with channel binding is enforced, relaying a captured hash becomes much harder. Extended Protection for Authentication (EPA) on web services such as AD CS web enrolment and Exchange closes the HTTP relay routes. Check what you actually enforce, not what the baseline document says.
When we test internal networks, unsigned SMB and NTLM-accepting web services are still among the findings we see most often. Exposure to a bug like this is less about Outlook and more about what a single hash can unlock afterwards. If you have not had that tested, an internal penetration test will show you the path from one leaked hash to domain access.
Detection
You are looking for two things: Outlook making unexpected outbound connections, and NTLM authentication to destinations that should never receive it. The queries below are starting points for Microsoft Defender for Endpoint advanced hunting. We have not yet tuned them against production data for this CVE, so expect to adjust them for your environment before turning them into alerts.
Outlook connecting out on SMB or WebDAV
// Classic Outlook making outbound SMB connections to public IPs.
// Legitimate traffic of this kind from outlook.exe should be rare.
DeviceNetworkEvents
| where Timestamp > ago(14d)
| where InitiatingProcessFileName =~ "outlook.exe"
| where RemotePort in (445, 139)
| where RemoteIPType == "Public"
| project Timestamp, DeviceName, InitiatingProcessAccountName, RemoteIP, RemotePort, RemoteUrl
| order by Timestamp desc
// When SMB is blocked, Windows can fall back to WebDAV over HTTP(S).
// The WebDAV client appears as rundll32 loading davclnt.dll; the command line
// usually contains the remote host. Correlate timestamps with Outlook activity.
DeviceProcessEvents
| where Timestamp > ago(14d)
| where FileName =~ "rundll32.exe" and ProcessCommandLine has "davclnt.dll"
| project Timestamp, DeviceName, AccountName, ProcessCommandLine
Likely false positives: legitimate internal file shares on public IP space, cloud file services that some organisations still map as WebDAV, and third-party Outlook add-ins that fetch content. Build an allow-list of known destinations rather than lowering the threshold.
NTLM leaving the network
On Windows, the Microsoft-Windows-NTLM/Operational log records outgoing NTLM authentication when NTLM auditing is enabled (Group Policy: Network security: Restrict NTLM: Outgoing NTLM traffic to remote servers set to Audit all). Event ID 8001 lists the target server name. Any target outside your own domains or known partners deserves a look.
Auditing outgoing NTLM is useful well beyond this CVE. It is also the first step towards restricting NTLM, because it shows you what would break.
Signs a captured hash was used
If you find a suspicious connection, assume the user's hash was captured and look for what came next:
- Sign-ins for that user from unfamiliar devices or locations shortly afterwards (Entra ID sign-in logs, on-premises event ID 4624 with logon type 3 and NTLM as the authentication package).
- NTLM authentication for that account on servers it does not normally touch, especially domain controllers and certificate services.
- Password spraying or successful logons against internet-facing services that still accept passwords alone.
Mitigation
In order of priority:
- Update Office everywhere classic Outlook runs. Push the November 2026 Office update on every channel you use, force an Outlook restart, and confirm the build afterwards. Report on Office build separately from Windows patch status.
- Block outbound SMB at the perimeter. Deny TCP 445 and NetBIOS (137–139) from user networks to the internet. There is rarely a business reason to allow it.
- Treat unexpected message files as suspicious. Until patching is complete, warn users about unsolicited
.msgor other Outlook item attachments, and consider blocking them at the mail gateway if your business does not rely on them. - Make relay harder. Enforce SMB signing on servers and domain controllers, LDAP signing and channel binding on domain controllers, and EPA on IIS-based services that accept Windows authentication.
- Protect the accounts that matter most. Put privileged users in the Protected Users group, which stops NTLM authentication for those accounts entirely. Test first: some legacy services will break.
- Plan to reduce NTLM. Use the audit data from the detection section to find what still depends on NTLM and move it to Kerberos. Microsoft has signalled for some time that NTLM is being phased out of Windows.
If you find evidence that the flaw was used against you, treat it as an identity compromise: reset the affected credentials, review sign-in activity for the period since the file was opened, and follow your incident response plan rather than handling it as a patching ticket.
What we know and what is unconfirmed
As of 11 November 2026, 09:00 UTC.
Stated in Microsoft's advisory and CISA's alert:
- CVE-2026-124194 is a spoofing vulnerability in Microsoft Office Outlook, fixed in the November 2026 security release on 10 November.
- Microsoft rates it Moderate, CVSS 6.5, and marks it as exploited.
- CISA added it to the Known Exploited Vulnerabilities catalogue on 10 November 2026, with a federal remediation deadline of 24 November 2026.
Widely reported by security press, not yet in the vendor text we have reviewed:
- Exploitation requires the user to open a crafted message file, after which Outlook connects to an attacker-controlled location without the external-content warning and leaks the user's NTLMv2 hash.
- The fix is delivered through Office updates and covers Microsoft 365 Apps and Office LTSC.
- Microsoft lists no workaround.
Unconfirmed or unknown:
- Who is exploiting it, against which sectors, and since when. No public attribution has been made.
- Whether the leak travels over SMB, WebDAV or both. Block and monitor both until this is clear.
- The exact list of fixed builds per update channel. Use Microsoft's release notes, not third-party summaries.
We will update this brief if Microsoft revises its advisory or credible reporting on the campaign appears.
Sources
- Microsoft Security Response Center, CVE-2026-124194 security update guide entry.
- CISA, CISA Adds Six Known Exploited Vulnerabilities to Catalog, 10 November 2026.
- CISA, Known Exploited Vulnerabilities catalogue.
- Microsoft Learn, Restrict NTLM: Outgoing NTLM traffic to remote servers.
Patching basics like these sit alongside the rest of the fundamentals in Ten Security Basics That Stop Most Real-World Attacks. If the detection side is where you are thin, In-House SOC or MSSP? covers how teams get round-the-clock eyes on queries like the ones above. Hunts like these are the day-one work of managed detection and response: an advisory lands, and the question is answered across the estate before the patch finishes rolling out.
Written by
Huzaifa Hafeez
Chief Technology Officer, FutaqSec