Back Shattered Outlook Exploit Rated Only 6.5, but Hashes Leak on Open
Microsoft says attackers are already abusing a flaw in Microsoft Office Outlook, and the fix has been available since November 10, 2026. The bug, tracked as CVE-2026-124194, is a spoofing vulnerability, and SecurityWeek reports that Microsoft says it is being actively exploited . Coverage from Yahoo Tech, Arabian Business, The Week and ET Now ran the story with a plain message for readers: install the update.
Microsoft classified the flaw as Moderate, with a reported CVSS score of 6.5. That label sounds mild. The impact is not. A successful attack can disclose a user’s NTLMv2 hash, which an attacker could then use to authenticate as that person. Our earlier look at the November Patch Tuesday release and its 1,094 fixes covered the raw numbers. This piece goes further into how the Outlook attack works, who needs to act, and which controls shrink the damage even on machines that patch late.
One caution up front. Some headlines say Microsoft told all Outlook users to update immediately. The advisory excerpts we reviewed do not contain that exact instruction. What they do contain is an exploited-in-the-wild flag, a clear attack description, and a shipping patch. That combination is reason enough to move fast, and we will separate what is confirmed from what is framing as we go.
Don't miss new tech stories on Google
Add Shattered once in the Google app and our stories show up in your news suggestions.
What Microsoft Says Is Being Exploited in Outlook
CVE-2026-124194 sits in Microsoft Office Outlook and belongs to the spoofing class. According to SecurityWeek, Microsoft confirmed active exploitation, which is the detail that separates this bug from the thousands of routine fixes in a typical monthly release. An exploited flaw means someone has already built a working attack and used it against real targets. The race between defenders and attackers has started before most organizations read the bulletin.
The attack needs two things from the victim’s side. An unauthenticated attacker sends a specially crafted message file, and the user has to be persuaded to open it. No credentials are required to start, and no prior foothold on the network is needed. Email delivery does the rest. That makes the entry cost low for an attacker with a decent phishing pretext, such as an invoice, a shared document, or a forwarded message from a supposed colleague.
Once opened, the file can make Outlook connect to a network location the attacker controls. Outlook normally warns users before it loads external content. In this case, the warning does not appear. The connection goes out quietly and carries authentication material with it. The table below summarizes the facts reported so far.
How the Outlook NTLMv2 Hash Leak Works
NTLM is a challenge-response authentication protocol that Windows has carried for decades. When a Windows client connects to a server that asks for authentication, the client proves who it is by computing a response from the user’s password-derived secret. The NTLMv2 hash is the product of that process. It is not the password itself, but it is close enough to be dangerous. Attackers can take it and try to crack it offline, or relay it to another service that accepts NTLM.
Why the missing warning matters
Outlook’s external-content warning exists to stop exactly this kind of silent outbound call. Remote images, linked resources and network paths in a message can reveal that a user opened a mail, and in a Windows environment they can also trigger automatic authentication. Microsoft’s description says the crafted file causes a connection without the usual warning. The user sees nothing that looks like a prompt, so the one human checkpoint in the chain is skipped.
What an attacker does with a hash
Microsoft says the attacker could use the disclosed hash to authenticate as the user. Two paths are common in practice. In the first, the attacker cracks the hash offline. Weak or reused passwords fall quickly, while long random ones hold up. In the second, the attacker relays the authentication attempt to another system in real time, which skips cracking entirely. Which path works depends on the target environment, including signing requirements on internal services and whether NTLM is still accepted at all.
Either route ends in the same place. The attacker holds a valid identity inside the organization without ever typing a password into a fake login page. That has real consequences for mailbox access, file shares and any internal web application that still trusts NTLM.
Who Is Affected and Which Updates to Install
Reports on the release name three product lines: Microsoft 365 for business, Office LTSC 2021 and Office LTSC 2024. The security updates for them shipped on November 10, 2026, as part of the monthly cycle. The reports we reviewed name only those three. Administrators running anything else should read Microsoft’s own entry in the Security Update Guide from the Microsoft Security Response Center , where each CVE gets its own advisory page, instead of assuming either way.
The split between the three products matters for patch speed. Microsoft 365 for business usually takes updates through its channel-based servicing, so many machines pull the fix without admin action once the channel releases it. LTSC builds behave differently. They sit on fixed servicing baselines, often in regulated or locked-down environments, and updates tend to wait for change windows. Those are the machines most likely to stay exposed the longest.
A practical order of operations for an IT team looks like this:
Confirm which Office builds run across the fleet, including virtual desktops and shared terminal servers.
Push the November 10 updates first to executives, finance, legal and IT staff, since they are the likeliest phishing targets.
Verify the installed build after the update instead of trusting a deployment dashboard.
Track LTSC machines separately, because they usually lag.
What Microsoft’s Own Description Says
Microsoft’s vulnerability text is short, and the wording carries most of the story. SecurityWeek quotes the company on how the attack starts and what it exposes. The first passage lays out the prerequisites and the credential risk in one go.
An unauthenticated attacker could send a specially crafted message file and convince a user to open it in an affected version of Microsoft Outlook. Opening the file could cause Outlook to connect to an attacker-controlled network location without displaying the usual external-content warning, potentially disclosing authentication information that could be used to impersonate the user. Microsoft, software and cybersecurity company, as quoted by SecurityWeek
An unauthenticated attacker could send a specially crafted message file and convince a user to open it in an affected version of Microsoft Outlook. Opening the file could cause Outlook to connect to an attacker-controlled network location without displaying the usual external-content warning, potentially disclosing authentication information that could be used to impersonate the user.
The same coverage records the impact statement. Microsoft’s description of CVE-2026-124194 says the following the credential at stake.
This vulnerability discloses a user’s NTLMv2 hash to the attacker who could use that to authenticate as the user. Microsoft, software and cybersecurity company, as quoted by SecurityWeek
This vulnerability discloses a user’s NTLMv2 hash to the attacker who could use that to authenticate as the user.
The Register also carried Microsoft’s wording and described the trigger in one line. It then recorded Microsoft’s account of what happens on the wire.
An attacker sends a specially crafted message file and convinces a user to open it. Microsoft, software and cybersecurity company, as quoted by The Register
An attacker sends a specially crafted message file and convinces a user to open it.
Outlook then connects to an attacker-controlled network location without showing the usual external-content warning, which discloses the user’s NTLMv2 hash. Microsoft, software and cybersecurity company, as quoted by The Register
Outlook then connects to an attacker-controlled network location without showing the usual external-content warning, which discloses the user’s NTLMv2 hash.
Read together, the four passages describe a clean chain. Delivery by email, a deliberate user action, a silent connection, and a stolen hash. Nothing in the text points to a memory-corruption bug or code execution. The flaw is a trust failure in how Outlook decides when to warn.
Headline Versus Advisory: Did Microsoft Say Update Immediately?
Several outlets framed the story as an urgent call to update. That framing is reasonable, and the facts support the urgency. It is still worth being precise what Microsoft itself has said, because the gap shapes how security teams write their own internal notices. Microsoft confirmed exploitation and published a fix. The advisory excerpts available to us do not contain a blanket instruction telling every Outlook user to update immediately.
The distinction is more than pedantry. Telling employees that Microsoft issued an emergency order when it did not can erode trust the time a real emergency arrives. Telling them that an exploited bug has a patch available, and that the patch should be installed this week, is accurate and just as persuasive. The table below sorts the claims circulating around this story.
Why a Moderate 6.5 Rating Understates the Risk
A CVSS score of 6.5 usually lands in the middle of a patch queue. Many vulnerability programs sort by score, and a Moderate flaw often waits behind Critical and High items. That habit is a problem here. CVSS measures the technical severity of a flaw in isolation. It does not weigh whether attackers are using the flaw today, and it cannot see how much damage a stolen hash does inside a particular network.
The user-interaction requirement pulls the score down. The attacker needs the victim to open a file, and that lowers the rating relative to a bug that fires with no clicks at all. Yet phishing campaigns clear that bar every day. A message file attached to a believable email gets opened often enough to pay off for the attacker. The interaction requirement slows an attack down by a little and does not stop it.
The better prioritization signal is the exploited flag. Security teams that follow a known-exploited approach, where active attacks jump a flaw to the front of the line, will treat this one as urgent regardless of the number. Teams that rely on score thresholds alone should make an exception. A Moderate label on an exploited credential-theft bug in the most widely used mail client is not a reason to wait for the monthly window.
The Preview Pane Question
Microsoft said the vulnerability cannot be triggered through the Preview Pane. That is good news with a limit. The Preview Pane lets a user glance at a message without fully opening it, and past Outlook flaws that fired on preview were especially nasty because a user did nothing at all. Here, the user must open the crafted file. The attack therefore depends on a deliberate action, and the passive-exposure scenario is off the table, according to Microsoft’s statement as reported.
The limit is behavioral. Users open message files all the time, because forwarded messages and saved emails are a normal part of office work. Training that says never click links does not cover a case where the file itself is the lure. Awareness messaging for this campaign should say something specific: treat unexpected message-file attachments the way you treat unexpected executables.
No reporting we reviewed says that reading a message in the normal reading view or previewing it fires the bug. Any headline implying otherwise goes beyond what Microsoft described. We treat that claim as unconfirmed.
One Fix Among 1,094 in the November Release
SecurityWeek reports that Microsoft’s November 2026 security updates addressed 1,094 vulnerabilities across its products. A release that large strains every patch team. Administrators have to test, stage and deploy a mountain of changes, and the single exploited Outlook bug can get buried in the volume. That is exactly why the exploited flag matters. It tells a team which one of the 1,094 deserves a fast lane.
Volume also changes the economics of attacker timing. When thousands of fixes land together, the window between release and widespread deployment tends to stretch. Attackers who already hold a working exploit for an Outlook bug have good reason to push campaigns hard in the days after Patch Tuesday, before enterprise rollouts finish. Defenders who pick out the Outlook update and ship it first shorten that window.
Readers who saw a different Outlook story this week should not confuse the two. The recent mail disruption that hit government agencies, covered in our report on the Outlook outage and its 130,000-plus Downdetector reports , is a separate event. Our follow-up on the second day of the Microsoft outage reported no cyber link to it. CVE-2026-124194 is a security flaw in the Outlook application, not an availability problem.
A Long History of Outlook Credential Leaks
This is not Outlook’s first brush with NTLM leakage. Security reporting has long tracked earlier flaws in the same family. CVE-2023-23397 involved malicious calendar invitations, and CVE-2024-21413 enabled NTLM credential theft through crafted Outlook links. The details differ, but the pattern repeats. A mail client reaches out to a network path an attacker chose, and Windows hands over authentication data on the way.
The repetition points at a design problem more than a string of unlucky bugs. Windows can send NTLM authentication automatically when a client connects to a resource that asks for it. Any code path in a mail client that fetches a remote resource becomes a possible credential leak unless something blocks the outbound request. Each patch closes one path. The underlying behavior stays until organizations restrict NTLM or block the traffic.
That history shapes the advice. Patching fixes this CVE. Hardening reduces the odds that the one, and there will be a one, does comparable harm. Security teams that treated the earlier Outlook bugs as one-off events and never touched their NTLM or SMB policies are in the same position again.
Hardening Steps Beyond the Patch
Reported mitigations for this class of bug include installing the November 2026 Office updates, blocking outbound SMB traffic, restricting or disabling NTLM where practical, and preventing direct outbound connections to untrusted network locations. None of these replaces the patch, but each one limits what a stolen or relayed hash can do. The table compares the main options on what they stop and what they cost.
The cheapest high-value step for many organizations is blocking outbound SMB, which normally runs over TCP port 445, at the network edge. Few workstations need to talk file-sharing protocols to the open internet. If the crafted message tries to pull in a remote path over SMB, the firewall can drop the attempt before any hash leaves the building. Verify the rule from inside the network, because misconfigured egress rules are common.
On the identity side, the long-term fix is less NTLM. Many organizations still allow it out of habit rather than need. Auditing which systems actually use NTLM, then restricting it in stages, is slow work but pays off across every future flaw of this type. Where NTLM must stay, enabling signing on internal services makes relay attacks harder. Teams unsure where to start can read our overview of how managed detection and response works against ransomware , since stolen credentials are a common first step in those intrusions.
How Security Teams Can Hunt for Signs of Abuse
Exploitation started before the patch, so patching alone does not tell a team whether it was hit. Hunting starts from the behavior Microsoft described. Outlook opens a message file, and shortly afterward the workstation tries to authenticate to an external address. Firewall and proxy logs are the first place to look. for outbound connections on SMB-related ports from user machines to public IP addresses in the weeks before the patch.
Authentication logs give the second signal. A hash that was stolen and relayed shows up as a login using a legitimate account from an unusual source or at an odd hour. Correlate those with users who recently received unsolicited message-file attachments. Mail gateway logs can list those attachments retroactively, which helps narrow the set of mailboxes to examine.
Teams that want to confirm what leaves a machine in a controlled test can capture traffic and filter for outbound authentication attempts. Our Wireshark display and capture filter lab walks through the filters needed. Run any such test only on machines and networks you own, and never with a live attack file from an untrusted source.
If a hunt finds a probable hit, treat the account as compromised. Reset the password, revoke active sessions, and review what the account touched since the suspected exposure. A reset also makes any offline-cracked hash worthless, which is the cleanest way to cut the attacker’s gain.
How This Compares With Classic Phishing
Classic credential phishing asks users to type a password into a fake page. Defenders have tools for that, including URL filtering, browser warnings and multi-factor authentication. The Outlook flaw skips the fake page. The credential material travels in the background, so the user never gets a chance to notice a suspicious login form. That makes user awareness weaker as a defense and makes technical controls matter more.
Multi-factor authentication helps less here than in a standard phishing case. A relayed NTLM exchange targets services that accept the hash-based protocol, and those services may not demand a second factor. Organizations should not assume their MFA rollout covers this path until they have checked which internal systems still accept NTLM.
Against that, the attack has a narrower footprint than a mass phishing wave. It needs the target to open a specific file type and needs a network path out. That gives defenders concrete choke points that a generic phishing campaign does not offer.
Enterprise and Market Impact
For most organizations the direct cost of this flaw is staff time. Patch testing, fleet verification, log hunting and possible password resets all land on small security and IT teams already working through a very large monthly release. Managed service providers carry the heaviest load because they patch many clients on different schedules. Our report on MSP ransomware recovery and client trust shows how a slow response can cost a provider customers.
The indirect effect is on how companies weigh Office servicing choices. LTSC builds suit locked-down environments, yet they trade speed for stability, and a flaw like this one exposes that trade. Expect IT leaders in regulated sectors to ask whether their change windows can handle an exploited bug within days instead of weeks. We found no market-moving financial data tied to this specific flaw, and we will not invent any.
For Microsoft, the story adds to a steady drumbeat of Outlook credential issues. The company shipped a fix on the normal monthly day and flagged the exploitation, which is what customers want to see. The harder question is architectural, namely how long Windows keeps sending NTLM authentication so readily. That question lives in the broader security coverage hub and in every enterprise architecture review that follows an incident like this one.
Five Predictions for the Weeks Ahead
These are our analysis, not reported facts, and each could turn out wrong.
More NTLM-leak variants will surface. Researchers who study Outlook and Office parsers tend to find sibling bugs after one is patched, so expect further reports of similar behavior.
Outbound SMB blocking will become standard guidance. After repeated Outlook episodes, more security baselines will treat blocked egress on SMB ports as a default.
Exploited status will outrank CVSS in more patch policies. A Moderate 6.5 flaw driving this much concern strengthens the argument for known-exploited prioritization.
Technical write-ups will appear. Once most fleets patch, security researchers usually publish analysis, which can lead to wider copycat use against laggards.
NTLM restriction projects will speed up. Budget owners who delayed NTLM audits will find this incident a useful argument to start them.
The takeaway: patch first, then harden. Install the November 10, 2026 Office updates on Microsoft 365 for business, Office LTSC 2021 and Office LTSC 2024. Verify the build on a sample of machines. Block outbound SMB at the edge, start auditing NTLM use, and brief staff on the specific lure, which is an unexpected message-file attachment.
users and small businesses on Microsoft 365 should confirm that Office has updated and restart Outlook if an update is waiting. Avoid opening message-file attachments from senders you did not expect, even if the Preview Pane looks harmless. Microsoft says the Preview Pane does not trigger the bug, and opening the file is the step that matters.
Frequently Asked Questions
What is CVE-2026-124194? It is a spoofing vulnerability in Microsoft Office Outlook that Microsoft says is being actively exploited. Microsoft rated it Moderate, with a reported CVSS score of 6.5, and fixes shipped on November 10, 2026.
What can an attacker steal with this Outlook flaw? Microsoft says the flaw can disclose a user’s NTLMv2 hash. An attacker could use that hash to authenticate as the user, either by cracking it offline or by relaying it to another service.
Do I have to open anything for the attack to work? Yes. Microsoft’s description says the attacker must persuade the user to open a specially crafted message file. The company also said the Preview Pane cannot trigger the flaw.
Which Office versions are affected? Reports name Microsoft 365 for business, Office LTSC 2021 and Office LTSC 2024. Check Microsoft’s Security Update Guide for the current list before assuming another version is safe or exposed.
Did Microsoft tell everyone to update immediately? Several outlets presented the story that way, and the urgency is justified by active exploitation. The advisory excerpts we reviewed do not contain that exact instruction, so we describe it as strong reporting framing rather than a quoted Microsoft order.
How does blocking SMB help? If the crafted file makes Outlook reach for a remote network path over file-sharing ports, an outbound block can stop the connection before authentication data leaves the network. It works as a safety net alongside the patch and does not replace it.
Is this connected to the recent Microsoft outage ? No link has been reported. Our coverage of the outage reported no cyber activity behind it, and CVE-2026-124194 is a separate flaw in the Outlook application. The vulnerability was part of the 1,094 fixes that SecurityWeek says Microsoft released in November 2026.
The full story
This article is one source in a clustered incident — the cluster page carries the summary, timeline and every other outlet covering it.
