Three vulnerabilities in Defender and BitLocker exploited after an uncoordinated disclosure

Microsoft’s Unexpected Vulnerability Disclosure: A Wake-Up Call for Cybersecurity

A researcher operating under the alias Nightmare Eclipse has recently made headlines by releasing exploit code for six Windows vulnerabilities between April and May 2026, bypassing the usual communication protocol with Microsoft. Among these, three significant flaws targeting Microsoft Defender and BitLocker have already been detected in active attacks. In a blog post dated May 27, the Microsoft Security Response Center condemned these uncoordinated disclosures as unjustifiable and announced that its Digital Crimes Unit would pursue involved parties. Notably, the BitLocker bypass remains unpatched, leaving organizations vulnerable until a manual mitigation can be applied.

The principles of coordinated disclosure hinge on a delicate balance between researchers who discover vulnerabilities and the vendors responsible for patching them. Typically, a researcher grants the vendor a grace period to develop a fix before making details public. This equilibrium functions effectively as long as both parties feel fairly treated. However, when trust erodes, the exploit code can find its way into the wild before a patch is released, transforming the gap between disclosure and the patch deployment into a risky exposure period for affected systems.

This breakdown in trust occurred at Microsoft in spring 2026. Under the alias Nightmare Eclipse, the researcher publicly shared proof-of-concept code for six vulnerabilities, named RedSun, UnDefend, BlueHammer, YellowKey, GreenPlasma, and MiniPlasma, between early April and mid-May. None of them were reported to the company prior to disclosure, and Microsoft has confirmed that three have been exploited in real-world attacks. The situation escalated on May 27 with a statement from the Microsoft Security Response Center, followed by the researcher threatening a new wave of disclosures on July 14.

Focus on Microsoft Defender and BitLocker

At the heart of this issue is Microsoft Defender, which is built into all versions of Windows by default. The flaw RedSun, tracked as CVE-2026-41091, allows for privilege escalation to SYSTEM rights due to an incorrect link resolution in the protection engine. Microsoft has classified it as “important,” assigning a CVSS score of 7.8, with exploitation already detected. A fix was rolled out in version 1.1.26040.8 of the engine on May 19. UnDefend, tracked as CVE-2026-45498, presents a denial-of-service vulnerability that freezes Defender’s signature update pipeline without triggering health alerts, ultimately degrading long-term protection. It has a CVSS score of 4.0 and was also detected being exploited, with a fix included in the antimalware platform 4.18.26040.7 released on the same day.

The most concerning vulnerability for organizations is YellowKey, tracked as CVE-2026-45585, which allows an attacker with physical access to the machine to bypass BitLocker’s encryption and access data on the system disk. This vulnerability impacts Windows 11 versions 24H2, 25H2, and 26H1, as well as Windows Server 2025. Unlike the two Defender vulnerabilities, YellowKey had no patch available at the time of publication. Microsoft issued a notice with a temporary mitigation in the form of a script that removes the executable autofstx.exe from the BootExecute registry value in the Windows recovery environment. Notably, a machine configured with a TPM and PIN is not exploitable, emphasizing an important consideration for encryption policies.

The sixth flaw, MiniPlasma, has not received an official CVE designation. The researcher describes it as a local privilege escalation in the cldflt.sys driver, capable of launching a SYSTEM shell on fully updated Windows 11 and Windows Server 2025 installations. BlueHammer, tracked as CVE-2026-33825, had been patched earlier in the year. The distinction between these vulnerabilities—registered, fixed, and merely mitigated—necessitates prioritization. RedSun and UnDefend emphasize the need for automated Defender update deployment, while YellowKey compels mobile fleet administrators to apply the mitigation script or enforce TPM and PIN configurations.

Breakdown Between Researcher and Vendor

The disclosures unfolded over a six-week span. The researcher first made BlueHammer public in April, explicitly claiming an uncoordinated disclosure. He then continued to release proof-of-concept code, initially hosted on GitHub and later on GitLab after facing the suspension of his GitHub account. By mid-May, he released MiniPlasma, alleging that a past fix for CVE-2020-17103 was no longer effective in the affected driver. He also claimed that Microsoft had silently patched RedSun initially without public acknowledgment.

The crux of the issue extends beyond the vulnerabilities themselves to the relationship between the parties involved. In a blog post, the researcher indicated that Microsoft had removed access to the Security Response Center account he used to report vulnerabilities and that he had not received any bounties for his previous disclosures. His attempts to seek clarification went unanswered, and he criticized the vendor for stating in its YellowKey advisory that there were violations of best practices for coordinated disclosure—a claim he considers defamatory. After his account was banned from GitHub on May 23, GitLab followed suit. Breaking his silence, the researcher set a deadline for July 14, promising a more significant disclosure and announcing an additional BitLocker bypass, dubbed Bitskrieg, along with collaboration with another researcher expected in June.

The true identity of the researcher remains uncertain. Several of his posts suggest a longstanding dispute with the vendor, though such claims cannot be verified. What is confirmed are the code releases, the CVE references issued by Microsoft, and the bans from platforms—the remainder constitutes the researcher’s claims made on his blog.

The Pressure on Coordinated Disclosure Standards

In its May 27 post, the Microsoft Security Response Center reaffirmed its commitment to coordinated disclosure, framing it as the industry standard, and indicated that it collaborates with hundreds of researchers yearly. The vendor asserted that uncoordinated disclosures that place exploit code into the hands of malicious actors targeting unpatched vulnerabilities are “never justifiable” and lead to real-world consequences. The announcement also indicated that its Digital Crimes Unit would continue pursuing legal action against these actors and those facilitating their criminal activities in collaboration with authorities, while maintaining that its reporting portal remains open to all, regardless of past interactions.

This firm stance has not garnered unanimous support in the security community; some experts argue that the threat of legal action and the banning of reporting accounts might dissuade well-meaning researchers rather than protect users. The conversation highlights the shared responsibilities inherent in coordinated disclosure. Vendors expect researchers to wait, while researchers seek timely responses, corrections, and acknowledgment of their efforts. When either party feels the other has failed to meet their obligations, the burden falls solely on goodwill, and a vulnerability window opens, endangering organizations deploying the affected products.

For a Chief Information Security Officer (CISO) or Chief Risk Officer (CRO) in Europe, the ramifications of this incident extend beyond Microsoft and directly impact vendor risk management. The security obligations imposed by European regulations, starting with NIS2 for essential and important entities, require control over the time between vulnerability discovery and patch deployment. When a major vendor and researcher engage in open conflict, this timeline becomes uncontrollable from the client’s perspective, pushing them to depend on their detection mechanisms, temporary mitigations, and monitoring of primary sources for security advisories. The promise of renewed disclosures by July 14 and an additional BitLocker bypass in June make it imperative for organizations to ensure the deployment of Defender updates and to tighten encryption practices for mobile devices, all while awaiting the outcome of this tense standoff.

Scroll to Top