Outlook zero-day fix: how to confirm it is installed beyond Windows Update
A Windows Update 'up to date' message may not cover Office, so the build number is the safer check.
Microsoft fixed an exploited Outlook flaw, CVE-2026-124194, on November 10, 2026. Secondary reports say the fix ships through Office updates rather than Windows Update, so the Office build number is the reliable check. The flaw can leak an NTLMv2 hash, which attackers can try to relay or crack.
Microsoft patched an Outlook flaw on November 10, 2026 that attackers were already exploiting. The bug, tracked as CVE-2026-124194, can make Outlook hand over a user's NTLMv2 authentication data after the user opens a crafted message file. A Windows Update screen reading 'You're up to date' does not by itself prove the fix is installed, because the patch is delivered as an Office update, according to two secondary reports.
That gap is why the practical test is the Office build number, not the Windows Update status. This article explains what the flaw does, how to compare a build against the fixed versions Microsoft lists, and what an attacker can and cannot do with a leaked NTLM hash. It is research context, not personal security advice, and the reporting behind it has real gaps that are spelled out below.
What the flaw is and what is known about the attacks
Microsoft titles the vulnerability 'Microsoft Outlook Spoofing Vulnerability.' According to the CVE record as mirrored by OpenCVE and threatint, the official description says the product's user interface fails to warn the user of unsafe actions, which lets an unauthorized attacker spoof over a network. The weakness class is CWE-356, the same mirror reports.
The word 'spoofing' can mislead. One source in the research notes that Microsoft uses the label because a stolen hash lets an attacker impersonate the victim, not because any email sender address is forged. The mechanism, as ThreatDefence describes it, is that a crafted message file opened in Outlook causes the program to connect to an attacker-controlled network location without the usual external-content warning. Microsoft's own wording, quoted by safeguard.sh and gblock.app, is that the bug discloses the user's NTLMv2 hash, which the attacker could use to authenticate as that user.
Microsoft rates the flaw Moderate, with a CVSS 3.1 base score of 6.5, according to Tenable. The scoring vector, as quoted by cve-security.com, is network-based, needs no privileges, requires user interaction, and has a high confidentiality impact with no integrity or availability impact. Microsoft's FAQ says the Preview Pane is not an attack vector, according to Rescana and Rapid7, so the user has to open the file.
Exploited before the fix, and a moderate score
Rapid7's table lists the exploitability status as 'Exploitation Detected' and says the flaw was not publicly disclosed before the patch. That is the basis for calling it a zero-day. Smtpedia and Shattered.io report that it was the only flaw Microsoft marked as exploited in the November release. Tenable's breakdown of the month's release shows 63 Critical, 1,029 Important, 1 Moderate and 0 Low ratings, and the single Moderate entry is this Outlook bug.
Shattered.io comments that this is the third month in a row in which the exploited bug was rated below Critical. That is the outlet's own observation. The practical point is that a 6.5 score and confirmed exploitation sit side by side, so the score alone is a poor guide to urgency here.
The CISA date: November 10, not November 12
Several outlets ran 'CISA confirms' stories on November 12, which may explain why that date circulates. The sources that state the listing date put it earlier. According to cve-security.com and SOC Prime, CISA added the flaw to its Known Exploited Vulnerabilities catalog on November 10, 2026, with a remediation due date of November 24, 2026, and listed known ransomware use as 'Unknown.' The reporting behind this article relied on those secondary sources and did not open CISA's entry directly.
The November 24 deadline binds US federal civilian agencies only, according to 1023jack.com, which adds that CISA urges other organizations to prioritize remediation as well. A separate Microsoft 365 outage began on November 12, but safeguard.sh says nothing published links it to this vulnerability.
Why Windows Update can say 'current' while Outlook is exposed
WindowsForum and Decryption Digest both report that the fix arrives through an Office update rather than through Windows Update. Neither is a Microsoft statement, and the reporting did not locate Microsoft's own explicit wording on the point. Microsoft's Office release notes do, however, list the fixed builds under Microsoft 365 Apps, which is consistent with a fix that is serviced by Office rather than by the operating system.
One verification gap matters. Safeguard.sh says the CVE is listed under Outlook on Microsoft Learn's security release notes. The excerpt reviewed for this article showed Outlook CVEs for November 10 but did not visibly include 124194, and it was truncated. The listing therefore could not be confirmed directly, and it is worth checking on the live page.
The reporting also flagged, as unverified background, that Click-to-Run Office installs update through their own updater on a channel schedule and use Windows Update only when a user or administrator has enabled that route. If correct, that would explain how an operating system can look fully patched while Office lags. It was not confirmed against Microsoft documentation here.
Which Office products are covered
The CVE record lists the affected products as Microsoft 365 Apps, Office 2021 and Office 2024. Rapid7 reports that Microsoft says all supported versions of classic Outlook receive patches. Whether the 'new Outlook' is affected, and whether older MSI-based Office such as 2016 or 2019 is covered, was not established in the sources reviewed. Readers on those products should consult Microsoft's advisory rather than assume either way.
How to check the build number
The fixed builds below come from Microsoft Learn's update history and download sizes pages for the November 10, 2026 release. The two volume-license rows come from the CVE record as mirrored by threatint and cve-security.com, which makes them somewhat weaker evidence than the Learn rows. Pouyam.com repeats the same volume builds but is not an independent source.
| Channel or edition | Version | Fixed build | Source |
|---|---|---|---|
| Current Channel | 2610 | 20527.20196 | Microsoft Learn |
| Monthly Enterprise Channel | 2610 / 2609 / 2608 | 20527.20194 / 20430.20264 / 20326.20320 | Microsoft Learn |
| Semi-Annual Enterprise Channel | 2610 / 2609 / 2608 | 20527.20192 / 20430.20262 / 20326.20318 | Microsoft Learn |
| Office LTSC 2024 | n/a | 16.0.17932.21140 | CVE record via threatint and cve-security.com |
| Office LTSC 2021 | n/a | 16.0.14334.21048 | CVE record via threatint and cve-security.com |
A working checklist
The build numbers above are verified against Microsoft Learn. The menu paths below are not. They were not checked against Microsoft documentation in this reporting, so the on-screen labels may differ by version.
- In Outlook, open File, then Office Account, then About Outlook, and read the line showing the version and build.
- Find the matching channel or edition in the table and compare the build. A build at or above the fixed number for that channel is patched.
- If the build is lower, the Office Account page typically offers an Update Options menu with an Update Now command.
- On a work-managed PC, ask the IT department instead of changing update settings, since administrators may control the channel and timing.
What an attacker can do with a leaked NTLM hash
'NTLM hash' is shorthand. What leaks is NTLMv2 authentication material, a challenge-response exchange sometimes called Net-NTLMv2. It is not the plain-text password and not necessarily the stored NT hash. Microsoft's own advisory language is 'NTLMv2 hash.' Two uses are documented: relaying it and cracking it.
Relay: borrowing the login in real time
According to Rapid7, the hash can be relayed live to impersonate the user against another service where that service lacks safeguards. Rapid7 names SMB or LDAP signing and Extended Protection for Authentication (EPA) for Exchange or Active Directory Certificate Services as the protections that block this. The SpecterOps white paper on NTLM relay describes the attacker's options the same way: relay the response to establish a session on a target server, or try to crack it.
Relay forwards an authentication exchange to another service. It does not hand the attacker a plain-text password, according to CyberPress, which also notes that the CVE description identifies spoofing rather than arbitrary code execution. The attacker's reach is therefore bounded by what the impersonated account can access and by whether the target service enforces signing or EPA.
Offline cracking: recovering the credential
Rapid7 says a motivated attacker can crack an NTLMv2 hash offline to recover the underlying credential. Success depends on how strong the password is. The research found no quantified figures on how often such cracking succeeds, so no rate can be given here.
Two single-source claims about persistence
ThreatDefence says a certificate obtained through relay to Active Directory Certificate Services can survive a password reset. That applies to enterprise environments, and no second source in the research repeats it. Pouyam.com says a captured hash derives from the old password, so a new password means the hash no longer matches. That is plausible for offline cracking, but it would not undo sessions or certificates an attacker had already obtained by relay.
The two claims illustrate why password changes alone may not close an incident in a corporate network. They are best read as hypotheses to confirm with Microsoft or an incident-response team, not as settled guidance.
Reducing exposure beyond the patch
Microsoft has not published a mitigation list specific to this CVE in the material reviewed. The closest official guidance concerns a similar Outlook flaw, CVE-2023-23397. In a March 2023 Microsoft Security Blog post, Microsoft advised updating Outlook, adding users to the Protected Users group, blocking outbound TCP 445 and disabling NTLM where possible. That is context for a different bug, not a verified recipe for this one.
The network route
The standard route for this kind of leak is outbound SMB on TCP 445. The 2023 Microsoft guidance says Windows falls back to WebDAV if SMB is unavailable. For the current flaw, ThreatDefence says blocking outbound SMB cuts off the attack, while noting WebDAV may remain a route. Microsoft has not confirmed which transport the November 2026 attacks used.
Restricting NTLM itself
Microsoft Learn documents that Windows 11 24H2 and Windows Server 2025 add an SMB client policy that can block NTLM for outbound connections, with exceptions that administrators can specify. A Microsoft Tech Community post on disabling NTLM warns against adding service or computer accounts to Protected Users. These are administrator-level controls, and a home user is unlikely to configure them directly.
Impact may also be concentrated in on-premises or hybrid Active Directory environments. CERT-EU said in a 2023 advisory about the similar bug that online services such as Microsoft 365 do not support NTLM authentication. Because that advisory predates this flaw, it is context only. No source in the research quantified the risk to home users with local or Microsoft-account logins.
Limits of the evidence
Most of the core facts rest on two underlying records, Microsoft's advisory and CISA's catalog entry, which many outlets repeat. Repetition is not independent confirmation. Rapid7 and Tenable add genuine expert analysis, but the reporting did not open the primary advisory pages directly.
Some inconsistencies are unresolved. CISA's assessment lists exploitation as active and one vector shows exploit code maturity as High, according to SecurityOnline, while another CVE mirror shows a different value. Public proof-of-concept code is tracked on GitHub, but one tracker notes it is untested against a vulnerable Outlook. The exploit prediction score is about 1.62% probability of exploitation in 30 days, according to thehackerwire. That score lags real events and does not measure the confirmed exploitation already reported.
What it means and what to watch
The episode shows how a patch can exist and still leave a machine exposed, because the update mechanism for Office is not always the one a user watches. The dependable check is a comparison of the installed build with the fixed build for the relevant channel. For managed PCs, that comparison belongs to the IT department, which controls update timing.
The leaked hash is serious but bounded. It can be relayed to services that lack signing or EPA, or cracked offline if the password is weak, and it does not by itself run code on a machine. Next to watch: whether Microsoft publishes indicators of compromise or details of the attacks, whether the November 24 federal deadline passes without reported incidents, and whether the Learn release notes and MSRC advisory confirm the points that secondary sources could only assert.
Frequently asked questions
Is Outlook patched if Windows Update says my PC is up to date?
Not necessarily. WindowsForum and Decryption Digest report that the fix comes through an Office update, not Windows Update. Microsoft's explicit statement on this was not located, so checking the Office build against the fixed build for the channel is the safer test.
Which Office build fixes CVE-2026-124194?
According to Microsoft Learn, the November 10, 2026 fix is Current Channel version 2610, build 20527.20196. Monthly Enterprise Channel builds are 20527.20194, 20430.20264 and 20326.20320. Semi-Annual Enterprise builds are 20527.20192, 20430.20262 and 20326.20318. The CVE record lists 16.0.17932.21140 for Office 2024 and 16.0.14334.21048 for Office LTSC 2021.
Can the flaw be triggered just by previewing an email?
Microsoft's FAQ says the Preview Pane is not an attack vector, according to Rescana and Rapid7. The user has to open the crafted message file for Outlook to connect to the attacker's location.
What can an attacker do with a leaked NTLMv2 hash?
Rapid7 describes two paths. The attacker can relay the hash to another service to impersonate the user where signing or Extended Protection is missing, or crack it offline to recover the credential. It is not a plain-text password, and the flaw itself is described as spoofing, not code execution.
Does changing my password help after a hash leak?
Pouyam.com says a captured hash derives from the old password, so a new password stops it from matching. That reasoning covers offline cracking but not sessions or certificates already obtained by relay, and ThreatDefence warns that a relayed certificate can survive a reset. Both claims are single-source, so people at risk should consult their IT or security team.
Sources (29)
- Update history for Microsoft 365 Apps (listed by date) - Office release notes , Microsoft Learn
- Download sizes for updates to Microsoft 365 Apps - Office release notes , Microsoft Learn
- Release notes for Microsoft Office security updates - Office release notes , Microsoft Learn
- Update History for Office Current Channel (Preview) - Office release notes , Microsoft Learn
- Guidance for investigating attacks using CVE-2023-23397 , Microsoft Security Blog (2023-03-24)
- SMB security hardening in Windows Server and Windows Client , Microsoft Learn
- Active Directory Hardening Series - Part 8: Disabling NTLM , Microsoft Tech Community
- Security Advisory 2023-018 , CERT-EU
- Patch Tuesday - November 2026 , Rapid7
- November 2026 Microsoft Patch Tuesday addresses 1,093 CVEs , Tenable
- Outlook CVE-2026-124194 NTLM Zero-Day: Patching and Hunting , ThreatDefence (2026-11-12)
- The Renaissance of NTLM Relay Attacks: Everything You Need to Know , SpecterOps
- Microsoft November 2026 Patch Tuesday: Windows Search, LDAP, and Schannel RCEs, and an Exploited Outlook Spoofing Flaw , Rescana
- CISA Confirms Active Exploitation of Microsoft Outlook Zero-Day Vulnerability , CyberPress (2026-11-12)
- CVE-2026-124194: Outlook Zero-Day Exposes NTLM Hashes , SOC Prime (2026-11-11)
- November 2026 Patch Tuesday: Exploited Outlook Zero-Day Requires Office Update , WindowsForum
- CVE-2026-124194 Outlook NTLM Hash Theft: Patch Now , Decryption Digest
- CVE-2026-124194 Outlook NTLM Zero-Day and the Microsoft 365 Outage , Safeguard.sh (2026-11-13)
- November 2026 Patch Tuesday: Outlook Zero Day Leaks NTLM Hashes , gblock.app (2026-11-11)
- CISA Confirms Active Exploitation of Outlook User Interaction 0-Day Vulnerability , Cryptika (2026-11-12)
- Outlook Zero-Day Under Attack: CISA Sets A November 24 Deadline. What To Patch Now , 1023 Jack
- Outlook Zero-Day CVE-2026-124194 Leaks NTLM Hashes: Patch Now , smtpedia
- Microsoft November 2026 Patch Tuesday: 1,094 CVEs , Shattered.io
- Outlook Zero-Day CVE-2026-124194: How to Patch (Nov 2026) , Pouyam.com
- Microsoft Outlook Zero-Day CVE-2026-124194 Exploited in the Wild Before November Patch , SecurityOnline
- CVE-2026-124194: Vulnerability Details - PoC Available , thehackerwire
- CVE-2026-124194 - Vulnerability Details - OpenCVE , OpenCVE
- CVE-2026-124194 , threatint
- CVE-2026-124194: Microsoft Office Outlook Missing User Interaction for Potentially Unsafe Action vulnerability , cve-security.com