All articles

NewsNovember 12, 20266 min read

Windows SignDrop: CLFS zero-day grants SYSTEM, no patch yet

SignDrop abuses CLFS log authentication and Windows Defender to gain SYSTEM privileges. Public exploit, no Microsoft patch yet: what to check now.

Shield with the word Security hit by blue light beams on a black background

Researcher Abdelhamid Naceri, known online as Nightmare Eclipse (Security Affairs also calls him Chaotic Eclipse), has published on GitHub the code and a technical write-up of SignDrop, a zero-day exploit for local privilege escalation on Windows. A standard user can obtain SYSTEM privileges by abusing the logic of a mitigation Microsoft introduced precisely to protect the Common Log File System (CLFS). At the moment there is neither a patch nor a CVE identifier.

BleepingComputer, The Hacker News, SecurityWeek and Security Affairs reported on it between 12 and 13 November 2026. The technical details that follow come from the researcher’s write-up, as reported by BleepingComputer, SecurityWeek and Security Affairs.

Background: CLFS log authentication

CLFS parses complex log files in kernel mode and has long been a frequent target for memory corruption vulnerabilities. To reduce the problem, Microsoft introduced CLFS log file authentication: according to Security Affairs, it was rolled out to the public channel from 28 October 2025, in two stages. For the first 90 days the system runs in learning mode; then full protection kicks in and every log file must have a valid authentication code, stored in a file with the .cnpf extension.

According to Naceri, attacks on CLFS dropped sharply after this mitigation, but the mitigation itself introduced the new problem: “Everything is done properly by Microsoft here but the logic allows abuse.”

How SignDrop works

The flaw is in the clfs.sys driver, in how it creates the .cnpf authentication file when a log file and its container are created:

  • when a process creates its own log files, CLFS correctly enforces access checks in the context of the calling process (with the OBJ_FORCE_ACCESS_CHECK flag, Security Affairs reports);
  • opening a log file also opens the associated container; if the container is empty, CLFS tries to sign it whether or not a signature file already exists;
  • opening a CLFS log file requires no special permissions: knowing its path is enough.

The chain described goes like this. The standard user prepares a log file that points to a container path, then forces Microsoft Defender to open it by calling the signature update RPC interface (ServerMpUpdateEngineSignature) and changing symbolic links in the Windows Object Manager. Defender opens the container in the context of the antimalware engine, follows the path to the associated file and, if the signature is missing, creates it with SYSTEM privileges; CLFS creates the new authentication file with the original file’s DACL, that is, with permissions chosen by the attacker. According to SecurityWeek, which has not tested the exploit, the proof of concept thereby creates a signature file in an attacker-controlled location, reopens it to write a payload and obtains a SYSTEM shell.

Naceri also claims the technique works against the self-defence (tamper protection) of any antimalware engine: the file creation appears to come from the product’s own legitimate process, so its drivers let it through, which would make it possible to turn off antivirus and EDR. The sources describe only the Defender chain in detail: the extension to other products is the researcher’s claim.

In the MITRE ATT&CK matrix the technique maps to T1068 (Exploitation for Privilege Escalation) and, for its use against self-defence, to T1562.001 (Impair Defenses: Disable or Modify Tools).

Affected systems

The researcher tested the exploit on Windows 11 26H2, Windows 11 26H2 Experimental and Windows Server 2025 Experimental, and estimates that around 80% of Windows installations worldwide are vulnerable; the estimate has not been independently verified.

The indications on which systems are exposed are not entirely consistent. According to Security Affairs, the problem is tied to the staged rollout: systems still in the learning period, or with the mitigation only partially enabled, are vulnerable. Naceri himself, quoted by SecurityWeek and Security Affairs, writes instead that an installation older than 90 days is “generally” vulnerable. The most precise check the researcher points to is the registry value HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\CLFS\Authentication\Mode: with 0 or 1 the system is vulnerable, with 2 or 3 it is protected (the latter as reported by Security Affairs).

According to SecurityWeek, the company Cyderes successfully compiled and tested the proof of concept and found no evidence of exploitation in real attacks. SecurityWeek asked Microsoft for comment, and the company had not responded at the time of publication.

Impact for European and Italian companies

SignDrop does not provide remote access: the attacker must already be able to run code on the machine as a standard user, for example after phishing or with stolen credentials. It is, however, the step that turns limited access into full control of the machine, and the claimed ability to turn off antivirus and EDR makes it attractive to anyone preparing a ransomware attack. With a public exploit and no patch, the time between publication and use in attacks can be short: according to Security Affairs, some exploits previously released by the same researcher were later exploited in real attacks. BleepingComputer notes that Microsoft has fixed only some of the zero-days Naceri has disclosed since April.

For NIS2 entities, this falls under handling vulnerabilities with no fix available: it should be assessed, documented and tracked until the patch, with compensating measures in the meantime.

What to do now

  1. Check the registry value CLFS\Authentication\Mode on workstations and servers, including through centralised management tools, and list the machines with a value of 0 or 1.
  2. Do not disable CLFS authentication: according to the researcher, it would remove SignDrop but reopen the door to the memory corruption vulnerabilities the mitigation blocks. For now, he says, there is no solution other than waiting for Microsoft’s patch.
  3. Do not run the proof of concept on production systems: the researcher suggests it as a check, but it is an exploit that opens a SYSTEM shell. If needed, use it only in an isolated lab.
  4. Follow Microsoft’s advisories and install the fix as soon as it is released, starting with the most exposed machines, such as user workstations and servers with interactive logons.
  5. Reduce initial access opportunities: least privilege, application control, multi-factor authentication.
  6. Keep an eye on protection: an antivirus or EDR that switches off or stops reporting, or a SYSTEM shell started by an unusual process, are signals to check immediately. The sources do not report specific indicators of compromise.

How ARXDOME helps

ARX Shield shows, for each machine, the status of the protection built into the operating system (health, definitions, detections, machines not reporting): protection that suddenly switches off stands out. When Microsoft releases the fix, ARX Shield’s vulnerability management helps you see which machines are still to be updated. ARX Sentry groups security events, including suspicious executions and privileged logons, and with ARX Shield you can isolate a suspicious machine from the network while you investigate.

Last updated:

Any questions?

Let’s talk about your company’s needs, with no obligation.

Contact us

Related articles