SignDrop: The Unpatched CLFS Flaw Giving Attackers the Keys to Windows

Red | Vulnerability
SignDrop: The Unpatched CLFS Flaw Giving Attackers the Keys to Windows | Hive Pro Threat Advisory
Date of publication
November 12, 2026
Admiralty code
A1
TA number
TA2026335
First seen
November 10, 2026
Affected products
Microsoft Windows systems with the CLFS authentication mitigation on, in enforced (Mode 0) or learning (Mode 1) mode (confirmed on Windows 11 26H2, 26H2 Experimental and Server 2025 Experimental)

Summary

SignDrop is an unpatched, locally exploitable zero-day in the Windows CLFS driver (clfs.sys) that abuses the CLFS authentication file (.cnpf) to elevate a low-privileged user to SYSTEM. Successful exploitation gives an attacker full control of the host: persistent implants, access to any data, and the ability to disable antivirus or EDR tamper protection.

Because it needs only local, low-privileged access, it turns any initial foothold from phishing or a compromised standard account into full system compromise. With no patch available and the flaw present in both learning and enforced modes, organizations should keep the mitigation on and apply compensating controls and heightened detection until a fix ships.

CVENameAffected product
UnassignedSignDrop (Windows CLFS Authentication Local Privilege Escalation Vulnerability)Microsoft Windows

Vulnerability Details

#1

The Common Log File System (CLFS) is a kernel logging subsystem exposed through clfs.sys and used by many Windows components to store transactional data in base log files (.blf). Because clfs.sys parses log structures that user mode can influence, it has been one of the most abused components for local privilege escalation. Microsoft responded with CLFS authentication: each log is paired with a .cnpf authentication file that lets the kernel reject a tampered log instead of trusting it.

#2

SignDrop is a local privilege escalation flaw in how clfs.sys creates the .cnpf file when a log is created through the \CLFS device, best classified as improper authentication (CWE-287) rather than a memory-safety bug. Two behaviors combine to defeat CLFS's access checks: a log can carry an extra container at a caller-chosen path, and CLFS creates and signs that container's authentication file whenever the log is opened; and opening an existing log needs only the correct path, not special privilege.

#3

A standard user stages a CLFS log and uses object-manager symbolic links to control where the .cnpf file is written, then coerces a privileged process into opening the log. In the proof of concept, Windows Defender is triggered through a legitimate local RPC call that requests a signature update. Because the engine runs as SYSTEM, CLFS creates the file as SYSTEM and applies the original file's DACL, so the attacker also controls the permissions on the SYSTEM-owned object.

#4

The result is an arbitrary file create-and-write primitive with SYSTEM privileges. The proof of concept uses it to replace a DLL loaded by a Microsoft shared component, running attacker-chosen code as SYSTEM. The coercion works against any antimalware or EDR engine: because the file operation originates inside the security product's trusted process, its kernel driver allows it, letting the attacker alter protected files and disable AV or EDR self-protection.

#5

Exposure is set by the host's CLFS authentication mitigation mode, not by the Windows build. Microsoft defines Mode 0 as enforced, 1 as learning, 2 as disabled by an administrator and 3 as disabled automatically by the system. Hosts are vulnerable whenever the mitigation is on, in learning or enforced mode, so reaching enforcement does not close the gap. Microsoft shipped the mitigation to the release channel on October 28, 2025, and the researcher estimates roughly 80% of Windows installations were vulnerable at disclosure.

#6

The researcher confirmed the issue on Windows 11 26H2, Windows 11 26H2 Experimental, and Windows Server 2025 Experimental; these are the tested builds, not the limit of what is affected. The source is a researcher-published proof of concept: there is no evidence of in-the-wild exploitation, no attributed threat actor, and no associated malware. A SYSTEM-level outcome, a low-complexity local vector, and no vendor patch still place it in the high-severity range, and the disclosure includes a host configuration check defenders can use to find vulnerable systems.

Affected products, CPE and CWE
CVE IDAffected productsAffected CPECWE ID
UnassignedMicrosoft Windows with the CLFS authentication mitigation on: Mode 0 (enforced) or 1 (learning)cpe:2.3:o:microsoft:windows:*:*:*:*:*:*:*:*CWE-287

Recommendations

  1. 01

    Verify CLFS Authentication Mitigation Mode

    On your Windows hosts, inspect the registry value HKLM\SYSTEM\CurrentControlSet\Services\CLFS\Authentication\Mode. Microsoft defines 0 as enforced, 1 as learning, 2 as disabled by an administrator and 3 as disabled automatically by the system. A value of 0 or 1 means the mitigation is on and the host is vulnerable; 2 or 3 means it is off, which avoids SignDrop but re-exposes older CLFS memory-corruption bugs. Use this check to inventory exposed systems and prioritize them for compensating controls.

  2. 02

    Keep the Mitigation Enabled Until Microsoft Patches

    Do not set the Mode value to 2 or 3 to avoid SignDrop. Those values turn the CLFS authentication mitigation off, which Microsoft advises against and which the researcher warns re-exposes CLFS memory-corruption bugs. No configuration change closes this flaw: keep the mitigation on, rely on the compensating controls and detections in this advisory, and deploy Microsoft's fix as soon as it is released.

  3. 03

    Restrict and Monitor Local Code Execution

    Because this vulnerability requires local, low-privileged execution as a prerequisite, enforce least privilege, remove unnecessary local administrator rights, and apply application control (such as WDAC or AppLocker) to limit the ability of unauthorized code to run. Reducing the opportunities for an attacker to gain an initial local foothold directly reduces the opportunity to chain into SYSTEM via this flaw.

  4. 04

    Harden and Monitor Endpoint Security Tooling

    Given the author's claim that the technique can disable antivirus or EDR tamper protection, ensure tamper protection is enabled and centrally monitored, and alert on any unexpected stop, configuration change, or tamper-protection state change for your security agents. Treat a sudden loss of EDR telemetry on a host as a high-priority investigation trigger.

  5. 05

    Await and Prioritize the Vendor Patch

    No official Microsoft patch for this specific flaw is referenced in the source at this time. Monitor Microsoft's security update channel and apply any corresponding update on an expedited schedule once released. Security concerns can be directed to Microsoft at secure@microsoft.com.

  6. 06

    Detect Anomalous CLFS and File-Write Activity

    Monitor for unusual creation of CLFS authentication files (.cnpf) and base log files (.blf), particularly those originating from non-standard or controlled paths, and for unexpected processes spawning SYSTEM-level command shells. Based on the proof of concept, pay specific attention to the creation or replacement of hid.dll in C:\Program Files\Common Files\microsoft shared\ink by the antimalware service or other SYSTEM processes, to CLFS log activity associated with Windows Defender signature-update operations, and to temporary working directories named with a SigDrop_ prefix. Correlate these events with low-privileged parent processes to surface potential privilege-escalation attempts.


Potential MITRE ATT&CK TTPs

Privilege Escalation
T1068Exploitation for Privilege Escalation
T1574Hijack Execution Flow
T1574.001DLL
Defense Evasion
T1562Impair Defenses
T1562.001Disable or Modify Tools
T1222File and Directory Permissions Modification
T1222.001Windows File and Directory Permissions Modification
Execution
T1059Command and Scripting Interpreter
T1059.003Windows Command Shell

References

References

What Next?

At Hive Pro, it is our mission to detect the most likely threats to your organization and to help you prevent them from happening.

Reduce real exposure. Not just vulnerability volume.