CVE-2026-124194 - Microsoft Office Outlook Missing User Interaction for Potentially Unsafe Action Vulnerability

CVE-2026-124194

Microsoft Outlook - NTLMv2 Leak via a Message File That Skips the External Content Prompt

What is Microsoft Outlook?

Outlook is the desktop email, calendar, and contacts client in Microsoft Office and Microsoft 365, and the default mail application on most corporate Windows endpoints. It reads and renders untrusted content all day, which is its job, and it runs in the logged-on user's Windows session with that user's credentials available to the operating system.

That second property is what makes Outlook a recurring source of credential-disclosure bugs. When Windows is asked to fetch a resource from a network location, it will often attempt to authenticate to that location automatically using the current user's credentials. If an attacker can get Outlook to reach out to a server they control, Windows hands over an NTLMv2 challenge-response for a domain account without the user doing anything deliberate. Outlook has guards against this, chiefly the prompt shown before loading external content, and the history of this bug class is largely a history of ways to skip that prompt.

Overview

CVE-2026-124194 is exactly that: a way to skip the prompt. An attacker crafts a message file and persuades a user to open it in Outlook. On opening, Outlook connects to an attacker-controlled network location without displaying its usual external content warning, and Windows authenticates to that location with the current user's credentials, disclosing an NTLMv2 challenge-response to the attacker's server.

Microsoft classifies it as a spoofing vulnerability, rates it Moderate, and scores it 6.5 under CVSS 3.1 (AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:N/A:N). The CWE is CWE-356, in Microsoft's framing a missing user interaction for a potentially unsafe action: the software performs something consequential on the user's behalf without the warning that was supposed to precede it.

The important detail is the exploitation marking. Microsoft shipped the fix in the November 2026 Patch Tuesday release on 2026-11-10, a release that addressed well over a thousand CVEs, and flagged this one as Exploitation Detected. It was not publicly disclosed before the patch, and it is the only CVE from that release that CISA added to the KEV catalog the same day, with a BOD 26-04 deadline of 2026-11-24. A Moderate rating and a 6.5 score understate its priority considerably.

Affected Versions

Product Vulnerable Fixed
Microsoft 365 Apps for Enterprise (32-bit and 64-bit) Builds before the November 2026 security update November 2026 Microsoft 365 Apps security update
Microsoft Office LTSC 2021 (32-bit and 64-bit) Builds before the November 2026 security update November 2026 Office security update
Microsoft Office LTSC 2024 (32-bit and 64-bit) Builds before the November 2026 security update November 2026 Office security update

The fix arrives through the Microsoft 365 Apps and Office update channels, not a Windows cumulative update. An endpoint fully patched for Windows in November is not necessarily patched for this.

Technical Details

The attack has three stages, and only the first is novel. First, a crafted message file causes Outlook to resolve and request a resource from a remote location the attacker specifies, and the code path taken bypasses the external content warning that normally gives the user a chance to refuse. Second, the Windows network stack, asked to reach that location, offers the current user's credentials and performs an NTLM exchange with the attacker's server. Third, the attacker keeps the resulting NTLMv2 challenge-response.

What that captured material is worth depends on the environment, and it is worth being precise because the shorthand "leaks your hash" oversimplifies. An NTLMv2 challenge-response is not a password-equivalent credential that can be replayed at will. It can be fed to an offline cracker, which succeeds or fails on the strength of the account's password. Alternatively it can be relayed in real time to another service that accepts NTLM and does not enforce signing or channel binding, which is how this class of bug typically turns into authenticated access. Neither outcome is automatic, and a tenant with strong passwords, NTLM restricted or disabled, and signing enforced gets much less out of this than one without.

Microsoft's FAQ states that the Preview Pane is not an attack vector, which is a meaningful limitation: the user has to open the file, not merely have it selected in the message list. There is no authentication requirement on the attacker's side, attack complexity is low, and no workaround has been published.

Discovery

Microsoft's acknowledgements for CVE-2026-124194 credit a researcher using the handle nono, and Benoit Sevens, whose credit line notes the use of Claude from Anthropic in the work. The flaw was privately reported, and Microsoft records it as not publicly disclosed before the patch.

Microsoft has not said whether the report came from analysis of an in-the-wild sample or from independent research that converged with attacker activity already underway. Given that the CVE arrived marked Exploitation Detected on the day of its first advisory, with no prior public disclosure, the former is the more likely of the two, but this is inference rather than a stated fact.

Exploitation Context

Microsoft's Exploitation Detected marking is the whole of the confirmed exploitation record. The company has not identified a threat actor, described the campaigns, named targeted sectors or regions, estimated how many organizations were affected, or published indicators of compromise. CISA lists known ransomware use as unknown.

The lack of detail should not be read as the attacks being minor. NTLM credential theft by crafted document or message is the opening move of a well-worn intrusion pattern, used by both espionage actors and ransomware affiliates, and it is attractive precisely because it needs no malware on the endpoint and generates traffic that looks like ordinary network activity. A single opened file yields material that is then cracked or relayed somewhere quieter.

One source derived from Tenable's data was still showing no recorded evidence of exploitation at the time of writing, which conflicts with both Microsoft's marking and the KEV listing. Treat that as a lag in a third-party feed rather than a genuine disagreement; Microsoft's own advisory is authoritative here.

Remediation

  1. Apply the November 2026 Office security update to Microsoft 365 Apps for Enterprise, Office LTSC 2021, and Office LTSC 2024, on both 32-bit and 64-bit installations. Confirm it landed by checking the Office build number rather than assuming the channel delivered it; Current Channel, Monthly Enterprise, and Semi-Annual all move on different schedules.
  2. Do not treat Windows patching as coverage. This fix ships through the Office update channels only, so an estate with strict Windows patching and loose Office patching is still exposed.
  3. Block outbound SMB (TCP 445) and unnecessary outbound authentication from user workstations at the perimeter. This does not fix the bug but it breaks the step where the credential actually leaves the network, and it is the single most effective compensating control while updates roll out.
  4. Reduce NTLM exposure generally. Restrict NTLM outbound to remote servers by Group Policy, enforce SMB signing and LDAP channel binding so captured responses cannot be relayed, and move services to Kerberos where possible. These controls blunt the entire bug class, not just this CVE.
  5. Hunt for the leak. Review egress and proxy logs for workstation connections to unfamiliar external hosts over SMB, WebDAV, or HTTP shortly after a user opened a message file, and correlate Outlook activity with authentication events where telemetry allows. No indicators of compromise have been published, so this is pattern hunting rather than a defined search.
  6. If evidence of exposure is found, reset the affected user's password and treat any service account whose credentials were available on that host as exposed too. Preserve logs and endpoint state before remediating, in line with CISA's Forensics Triage Requirements under BOD 26-04.
  7. Federal civilian agencies must remediate or discontinue use by 2026-11-24. The Moderate severity rating should not drive the schedule here: confirmed exploitation and a credential-theft outcome should.

Key Details

PropertyValue
CVE ID CVE-2026-124194
Vendor / Product Microsoft - Office
NVD Published2026-11-10
NVD Last Modified2026-11-10
CVSS 3.1 Score6.5
CVSS 3.1 VectorCVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:N/A:N
SeverityMEDIUM
CWE CWE-356 find similar ↗
CISA KEV Added2026-11-10
CISA KEV Deadline2026-11-24
Known Ransomware Use No

CVSS 3.1 Breakdown

Attack Vector
Network
Attack Complexity
Low
Privileges Required
None
User Interaction
Required
Scope
Unchanged
Confidentiality
High
Integrity
None
Availability
None

Required Action

CISA BOD 26-04 Deadline: 2026-11-24. Apply mitigations in accordance with vendor instructions, ensuring compliance with CISA's BOD 26-04 Prioritizing Security Updates Based on Risk (see URL in Notes) guidance and CISA's "Forensics Triage Requirements" (see URL in Notes). Follow applicable BOD 26-04 guidance for cloud services or discontinue use of the product if mitigations are unavailable. Stakeholders are responsible for evaluating each asset's internet exposure and ensuring adherence to BOD 26-04 patching guidelines.

Timeline

DateEvent
2026-11-10Microsoft publishes the advisory in the November 2026 Patch Tuesday release and marks the CVE Exploitation Detected
2026-11-10CVE published; NVD scores it 6.5 under CVSS 3.1
2026-11-10Added to CISA Known Exploited Vulnerabilities catalog
2026-11-24CISA BOD 26-04 remediation deadline