November 2026 Patch Tuesday: Outlook Zero-Day and Exchange RCE

Microsoft and Adobe released massive November 2026 security updates patching over 1,000 vulnerabilities, including an exploited Outlook NTLM hash disclosure zero-day and a zero-click Exchange Search RCE.

TL;DR

  • Active Exploitation: CVE-2026-124194 is an actively exploited Microsoft Outlook spoofing flaw that bypasses external-content warnings to leak user NTLMv2 hashes.
  • Zero-Click Perimeter RCE: CVE-2026-116928 allows unauthenticated remote code execution on Exchange Servers via automated indexing of malicious attachments by the Windows Search component.
  • Massive Volume: Microsoft patched over 1,000 vulnerabilities (1,235 per Qualys audit, ~1,090 per ZDI count), including 63 Critical-severity bugs across Windows, Hyper-V, LDAP, and Schannel.
  • Cloud Advisory Revisions: Microsoft updated the advisory for Azure OpenAI privilege escalation flaw CVE-2026-119990, clarifying that no customer action is required.
  • Third-Party Ecosystem: Adobe addressed 110 CVEs across nine security bulletins, headlined by 24 CVEs in Adobe Connect (top CVSS 10.0) and 14 in Adobe Experience Manager (top CVSS 9.9).

The patch tuesday november 2026 security updates represent one of the highest-volume patch deployments in enterprise IT history, capping off back-to-back record release cycles. Security engineering teams face more than 1,000 resolved vulnerabilities from Microsoft—including 63 rated Critical—alongside nine comprehensive security bulletins from Adobe covering 110 flaws. Managing this update load requires precise, risk-based prioritisation rather than indiscriminate deployment.

Engineering teams must immediately address two distinct threat vectors: an actively exploited Outlook spoofing vulnerability (CVE-2026-124194) that coaxes authentication hashes over the network, and a zero-click remote code execution bug (CVE-2026-116928) in the Windows Search component triggered automatically by Exchange Server email indexing.

While overall patch volume remains elevated following the October release cycle (as documented in our November 2026 Patch Tuesday analysis of PHP2Shell and Brotli), defensive efforts must focus on authentication coercion vectors, core domain infrastructure like Active Directory LDAP, and public-facing application servers.

Executive Summary & November 2026 Threat Environment

According to a Qualys audit, Microsoft resolved 1,235 vulnerabilities in the 10 November 2026 release, while a parallel Zero Day Initiative review indexed ~1,090 unique CVEs. The release contains 63 Critical-severity flaws and 1,029 Important-severity vulnerabilities. Additionally, Microsoft delivered updates for 142 vulnerabilities within Chromium-based Microsoft Edge.

This drop follows an unprecedented October patch cycle that exceeded 1,400 CVEs. Managing successive monthly cycles containing over 1,000 vulnerabilities introduces severe testing and operational friction for enterprise systems administration. Defensive teams must separate high-risk perimeter, domain-level, and user-coercion vectors from lower-risk local bugs to prevent system destabilisation while maintaining protective coverage.

Vendor / Category Total Vulnerabilities Critical Severity Zero-Day / Active Exploits Top Priority Target Components
Microsoft Windows & Enterprise Services 1,235 (Qualys) / ~1,090 (ZDI) 63 1 (CVE-2026-124194) Outlook, Exchange / Windows Search, LDAP, Schannel, Hyper-V
Adobe Enterprise & Creative Suite 110 (across 9 bulletins) 42 (Connect & AEM combined) 0 Adobe Connect (APSB26-191), Adobe Experience Manager (APSB26-189)

Automated patch-diffing tools allow threat actors to reverse-engineer security fixes rapidly. Consequently, the window between patch publication and weaponisation against self-hosted infrastructure continues to collapse. Prioritising patches based purely on raw CVSS scores is insufficient; teams must evaluate exploitability, network exposure, and system role context.

Active Exploitation Analysis: CVE-2026-124194 Outlook Zero-Day

The single actively exploited flaw in the November release is CVE-2026-124194, an unauthenticated Microsoft Outlook spoofing vulnerability. Although assigned a moderate CVSS rating of 6.5 due to the requirement for user interaction, its exploitation in the wild elevates it to an immediate remediation priority.

Mechanism of Exploitation

The flaw centres on how Microsoft Outlook processes specially crafted message files (`.msg` or similar object encodings). An attacker crafts a malicious message file and delivers it via email or a shared location, convincing a target user to open it. Upon opening the file, Outlook executes out-of-band network parsing logic to retrieve external assets or definitions.

Warning Prompt Bypass

Standard Outlook implementations trigger an explicit warning prompt when a file attempts to fetch external content or initiate out-of-band connections. CVE-2026-124194 bypasses this prompt completely, initiating outbound connections without user notification.

When Outlook executes the out-of-band request to an attacker-controlled remote location (typically via Server Message Block / SMB using UNC paths like \\attacker-host\share), Windows automatically attempts to authenticate using the current user’s security context. This process exposes the user’s NTLMv2 hash over the network.

[Attacker System] <--- Sends Malicious .msg File --- [Victim System]
                                                          |
[Victim User] Opens File (No Warning Prompt Shown) -------+
                                                          |
[Victim Outlook] Forces Outbound UNC/SMB Connection -----> [Attacker SMB Listener]
                                                            (Captures NTLMv2 Hash)

Impact and Authentication Relay Attack Paths

Once an attacker captures an NTLMv2 hash, two operational attack paths open:

  1. Offline Password Cracking: If the user account uses a weak password, the hash can be brute-forced offline to recover plain-text credentials.
  2. NTLM Relay Attacks: If the target enterprise has not enforced SMB Signing, Extended Protection for Authentication (EPA), or restricted NTLM usage, the attacker can relay the authentication response in real-time to internal services (such as Active Directory Web Services, Exchange Web Services, or internal SMB shares) to impersonate the target user.

Microsoft confirms that the Outlook Preview Pane is not an attack vector for CVE-2026-124194. The target user must explicitly open the file for the authentication coercion code path to trigger. Specific threat actor groups actively exploiting this bug in targeted campaigns have not been publicly disclosed by vendors or intelligence sources.

While CVE-2026-124194 requires user interaction, CVE-2026-116928 presents an unauthenticated, zero-click remote code execution risk rated CVSS 9.8. The underlying vulnerability resides within the Microsoft Windows Search component, but its integration with Microsoft Exchange Server creates an exposed perimeter attack surface.

Architecture of the Vulnerability Path

Microsoft Exchange Server relies on background indexing services to allow users to search mailboxes, calendars, and attachments quickly. When an incoming email arrives containing an attachment, Exchange automatically passes the attachment file to the underlying Windows Search component for content extraction and indexing.

CVE-2026-116928 is an out-of-bounds write flaw in the Windows Search component. An unauthenticated remote attacker can send a specially crafted attachment inside an email to any valid address on an exposed Exchange Server. The target mailbox recipient does not need to open the email, view the attachment, or interact with Outlook in any way.

[Unauthenticated Attacker]
       |
       v (Sends Email with Malicious Attachment over SMTP)
[Microsoft Exchange Server]
       |
       v (Automatic Attachment Background Indexing)
[Windows Search Component] --- Trigger CWE-787 Out-of-Bounds Write ---> [RCE in Service Context]

When Exchange invokes the Windows Search subsystem to index the incoming payload, the malformed structures in the attachment trigger the out-of-bounds write condition. This permits memory corruption and execution of arbitrary code within the security context of the indexing service on the underlying host platform.

Perimeter Isolation and Immediate Workarounds

Because Exchange Servers routinely face the internet to process inbound SMTP traffic, this flaw transforms a local indexer bug into a zero-click perimeter remote code execution vector. Organizations unable to apply the November cumulative update immediately should consider temporary containment measures:

  • Restrict or temporarily disable attachment indexing rules in Exchange Server settings where operational policies allow.
  • Enforce strict perimeter transport rules to block untrusted or complex file extensions at the email gateway before payloads reach Exchange servers.
  • Isolate Exchange Servers from internal management zones using network segmentation to prevent lateral movement if a service worker process is compromised.

Domain Infrastructure Risks: LDAP and Schannel Vulnerabilities

Core infrastructure components within Windows Server received critical fixes addressing unauthenticated network code execution threats targeting Active Directory Domain Controllers and basic transport security components.

CVE-2026-119958: Windows LDAP Out-of-Bounds Write

CVE-2026-119958 addresses an out-of-bounds write vulnerability (CWE-787) in the Windows Lightweight Directory Access Protocol (LDAP) service. An unauthenticated attacker can execute arbitrary code over the network by transmitting crafted LDAP requests to an exposed LDAP server or Domain Controller.

Because LDAP operates as a core identity mechanism across Active Directory environments, exploitation of this vulnerability on a Domain Controller grants full administrative execution context over the directory topology. No user interaction or prior domain authentication is required. Domain Controllers exposed directly to untrusted networks or hybrid cloud tunnels must be patched immediately.

CVE-2026-119952: Windows Schannel Remote Code Execution

Windows Secure Channel (Schannel) is the security package that implements the TLS and SSL authentication protocols in Windows systems. CVE-2026-119952 is caused by an observable timing discrepancy within Schannel’s cryptographic processing logic, enabling remote code execution.

Exploitation Probability Rating

Microsoft assigned CVE-2026-119952 an assessment of “Exploitation More Likely,” despite rating its attack complexity as High. High-complexity ratings often depress overall CVSS base scores, but Microsoft’s internal exploitability indexing signals elevated operational risk.

Schannel vulnerabilities affect any Windows server listening for incoming TLS connections, including Web Servers (IIS), Remote Desktop Gateways, and enterprise application ports. When deploying updates for LDAP and Schannel, operations teams must validate protocol compatibility in staging domains to prevent unintended disruptions to legacy authentication clients or third-party TLS handshakes.

Cloud and Virtualisation Flaws: Azure OpenAI, Hyper-V, and VFP

The November cycle includes fixes across virtualisation hypervisors, cloud management services, and core virtual networking stacks.

Azure OpenAI Vulnerability (CVE-2026-119990) Guidance

Initial release documentation for CVE-2026-119990 (an elevation of privilege flaw caused by missing authentication for a critical function in Azure OpenAI) listed the issue as requiring customer action. However, Microsoft subsequently revised the advisory to state that no customer action is required.

The underlying infrastructure changes have been applied managed-side by Microsoft. Organizations utilizing Azure OpenAI services do not need to apply host patches or adjust tenant deployment configurations for this vulnerability.

Windows Hyper-V Privilege Escalation Vectors

Microsoft patched several high-severity local privilege escalation flaws in Windows Hyper-V that allow authenticated users inside a guest virtual machine to execute code or elevate privileges on the underlying host operating system or host kernel context:

  • CVE-2026-119960: Privilege escalation resulting from the use of an uninitialised resource within Hyper-V memory management routines.
  • CVE-2026-113865: A Time-of-Check to Time-of-Use (TOCTOU) race condition flaw enabling guest-to-host privilege escalation.
  • CVE-2026-119947, CVE-2026-113860, CVE-2026-113861, CVE-2026-113864: Use-after-free (UAF) vulnerabilities within Hyper-V host management interfaces.

Enterprise cloud providers, multi-tenant virtual desktop infrastructure (VDI) operators, and hyper-converged clusters must update host hypervisors to protect virtual machine boundaries.

CVE-2026-95962: Windows Virtual Filtering Platform Network RCE

The Windows Virtual Filtering Platform (VFP) is a core component that provides network packet filtering, virtual switch extensions, and network isolation for Hyper-V and software-defined networking (SDN) stacks. CVE-2026-95962 represents a use-after-free flaw in VFP packet handling.

An unauthenticated network attacker can trigger remote code execution on the host system by sending specially crafted network packets to a vulnerable VFP instance. Although exploitation requires racing a internal memory state condition, this flaw presents severe risk to virtualised network infrastructure and tenant isolation models.

Enterprise Software Ecosystem: Adobe Security Bulletins Breakdown

Adobe released nine security bulletins addressing 110 unique vulnerabilities. While none are currently reported as under active exploitation in the wild, two enterprise server bulletins require immediate deployment priority.

Bulletin ID Product Total CVEs Highest Severity Top CVSS Score Deployment Priority
APSB26-191 Adobe Connect 24 Critical 10.0 Priority 2
APSB26-189 Adobe Experience Manager 14 Critical 9.9 Priority 2
APSB26-102 Adobe Audition 11 High 8.7 Priority 3
APSB26-100 Adobe After Effects 17 High 8.6 Priority 3
APSB26-190 Adobe Dimension 1 High 8.2 Priority 3
APSB26-181 Adobe Photoshop 13 High 7.8 Priority 3
APSB26-101 Adobe Premiere 11 High 7.8 Priority 3
APSB26-103 Adobe Media Encoder 6 High 7.8 Priority 3
APSB26-184 Content Credentials SDK 13 Medium 6.5 Priority 3

Critical Server Bulletins: Adobe Connect and AEM

According to details in the Zero Day Initiative review, bulletin APSB26-191 for Adobe Connect is the most severe non-Microsoft update this month. It covers 24 CVEs, 18 of which are rated Critical. The maximum CVSS score is 10.0. Each of the CVSS 10.0 vulnerabilities is reachable over the network without authentication or user interaction, and causes a scope change (Scope: Changed).

Bulletin APSB26-189 addresses 14 CVEs in Adobe Experience Manager (AEM), including four Critical bugs rated CVSS 9.9. These comprise two incorrect authorisation flaws and two arbitrary code injection vulnerabilities. Exploitation allows unauthenticated or low-privileged network attackers to achieve arbitrary code execution on host AEM instances.

Desktop Media Processing Tools

The remaining Adobe bulletins address file parsing vulnerabilities (such as heap-based buffer overflows and use-after-free bugs) across desktop applications including Photoshop, Premiere, After Effects, Audition, Dimension, Media Encoder, and the Content Credentials SDK. These flaws typically require a user to open a maliciously crafted image, audio, or video file. Enterprise administrators should deploy these client updates during standard application patching schedules.

Practitioner Remediation Roadmap and Prioritisation Matrix

To manage the volume of updates safely across production systems, operations engineering teams should execute a phased deployment model based on asset exposure and vulnerability mechanics.

Phased Rollout Strategy

  1. Phase 1: Zero-Day & Zero-Click Perimeter (Target: 0–24 Hours)
    • Deploy Windows cumulative updates to Microsoft Exchange Servers and systems hosting the Windows Search component (CVE-2026-116928).
    • Deploy Office/Outlook updates across client workstations to address CVE-2026-124194.
    • Update self-hosted Adobe Connect (APSB26-191) and Adobe Experience Manager (APSB26-189) instances.
  2. Phase 2: Core Infrastructure & Hypervisors (Target: 24–72 Hours)
    • Patch Active Directory Domain Controllers to address LDAP RCE (CVE-2026-119958).
    • Patch Windows servers hosting public TLS services to resolve Schannel RCE (CVE-2026-119952).
    • Deploy host updates across Hyper-V clusters and VFP-enabled network hosts (CVE-2026-119960, CVE-2026-95962).
  3. Phase 3: General Fleet & Desktop Ecosystem (Target: Within 7 Days)
    • Roll out cumulative updates across general enterprise Windows endpoints and member servers.
    • Update remaining Adobe desktop tools (Photoshop, Premiere, After Effects) via central package management.

Defensive Hardening and Mitigation Configurations

In parallel with patch deployment, security operations teams should enforce network-level mitigations to neutralize authentication coercion vectors. The primary defense against NTLM leakage (such as CVE-2026-124194) is blocking outbound SMB traffic at the perimeter firewall.

Execute the following PowerShell command on endpoint management policies to audit or restrict outbound SMB traffic over port 445 (test in staging environments prior to fleet deployment):

# Query current outbound SMB client configuration
Get-SmbClientConfiguration | Select-Object -Property EnableBandwidthThrottling, EnableSecuritySignature, RequireSecuritySignature

# Block outbound SMB (TCP 445) to untrusted remote networks via Windows Defender Firewall
New-NetFirewallRule -DisplayName "Block Outbound SMB Port 445" `
                    -Direction Outbound `
                    -Action Block `
                    -Protocol TCP `
                    -RemotePort 445 `
                    -Profile Any

To verify that Exchange Server attachment indexing does not leak operational exceptions during patching cycles, administrators can inspect event logs for Windows Search service faults using PowerShell:

# Search event logs for Windows Search Service crash events (Event ID 1000 / 7034)
Get-WinEvent -FilterHashtable @{
    LogName = 'Application', 'System'
    ProviderName = 'Application Error'
    StartTime = (Get-Date).AddDays(-3)
} | Where-Object { $_.Message -match "SearchIndexer.exe" } | Format-List -Property TimeCreated, Message

Common Operational Mistakes

  • Deprioritising Low-CVSS Bugs: Assuming CVE-2026-124194 can be delayed because its CVSS base score is 6.5. Because it is actively exploited in the wild, its mitigation priority supersedes unexploited higher-scoring vulnerabilities.
  • Neglecting Outbound SMB Filtering: Relying solely on endpoint patches while leaving outbound TCP port 445 wide open to the internet allows future, unpatched NTLM coercion vectors to succeed.
  • Skipping Host Updates on Virtualisation Clusters: Patching guest virtual machines while delaying Hyper-V host updates leaves hypervisors exposed to guest-to-host privilege escalation and VFP network packet corruption.

Key Takeaways and Architectural Implications

The November 2026 patch cycle reinforces key operational lessons for security architects and engineering leaders:

  • Authentication Coercion Requires Egress Controls: Vulnerabilities like CVE-2026-124194 succeed because corporate environments permit outbound SMB connections to arbitrary internet hosts. Strict egress filtering on TCP port 445 is an essential baseline control.
  • Shared Components Expand Attack Surfaces: The zero-click exploit path of CVE-2026-116928 demonstrates how integrating internal OS components (Windows Search) into server pipelines (Exchange Server) can expose local bugs to unauthenticated remote attackers over the internet.
  • Distinguish Managed PaaS from Self-Hosted Roles: As seen with the Azure OpenAI advisory revision (CVE-2026-119990), enterprise security teams must clearly differentiate cloud-vendor patching responsibilities from self-hosted assets to avoid unnecessary deployment overhead.

Executive Checklist for Engineering Leads

  • Verify perimeter firewalls block outbound SMB (TCP 445) traffic for all enterprise subnets.
  • Confirm 100% patch deployment for Exchange Servers targeting Windows Search (CVE-2026-116928).
  • Deploy Windows cumulative updates to Active Directory Domain Controllers (CVE-2026-119958).
  • Audit Adobe Connect and Adobe Experience Manager instances for APSB26-191 and APSB26-189 patch status.

Frequently asked questions

Is the Outlook zero-day (CVE-2026-124194) exploitable via the Outlook Preview Pane?

No, vendor analysis confirms that the Outlook Preview Pane is not an attack vector for CVE-2026-124194. Exploitation requires a user to open the specially crafted message file explicitly. However, once opened, the file suppresses standard external-content warning prompts and automatically initiates an out-of-band network connection to leak the user’s NTLMv2 hash.

Why is CVE-2026-116928 considered zero-click if it requires an email attachment?

CVE-2026-116928 is classified as zero-click because Microsoft Exchange Server automatically processes and indexes incoming email attachments via the background Windows Search component upon arrival. The target recipient does not need to open the email, view the attachment, or interact with an email client for the underlying out-of-bounds write vulnerability to trigger.

Do enterprise administrators need to perform manual remediation steps for Azure OpenAI vulnerability CVE-2026-119990?

No, enterprise administrators do not need to take action. Although initial Microsoft release documentation indicated customer action was required for CVE-2026-119990, Microsoft subsequently updated the advisory to clarify that no customer remediation is necessary. The fix was implemented directly on the cloud platform by Microsoft.

Which Adobe product update represents the highest vulnerability severity in November 2026?

Adobe Connect bulletin APSB26-191 represents the highest vulnerability severity in November 2026. It addresses 24 total vulnerabilities, 18 of which are rated Critical with a maximum CVSS score of 10.0. These flaws permit unauthenticated remote code execution over the network without user interaction.

How does outbound NTLM blocking mitigate authentication coercion attacks like CVE-2026-124194?

Outbound NTLM blocking (specifically restricting SMB TCP port 445 outbound at perimeter firewalls and enforcing Windows NTLM restriction policies) prevents the victim’s client from establishing out-of-band SMB connections to untrusted external servers. Even if a user opens a malicious file, the network block prevents the host from sending NTLM authentication hashes to the attacker’s listener.

Sources and further reading

This article was drafted with AI assistance from the publicly cited sources and passed automated checks for structure and sourcing. Verify critical details against the vendor advisories linked above before acting.