Researchers at Cyera have disclosed a chain of four vulnerabilities in OpenClaw, an open-source autonomous AI agent platform that Nvidia and Tencent have built enterprise products on top of. The chain - CVE-2026-44112 (CVSS 9.6), CVE-2026-44113, CVE-2026-44115, and CVE-2026-44118 - lets an attacker who can influence the agent's input (through a malicious plugin, prompt injection, or compromised tool output) break out of the OpenShell sandbox, read environment-stored API keys, elevate to owner-level privileges, and write persistent backdoors. Each step looks like normal agent behavior. Shodan and Zoomeye between them counted 65,000 to 180,000 public OpenClaw instances earlier in May. All flaws are fixed in OpenClaw 2026.4.22.
A researcher who goes by Chaotic Eclipse has dropped working proof-of-concept code on GitHub for a Windows local privilege escalation that gives SYSTEM access on fully patched Windows 11 Pro and Windows Server 2025. The bug lives in the Cloud Filter driver cldflt.sys and is, the researcher says, the same flaw Google Project Zero reported to Microsoft as CVE-2020-17103 in 2020, which Microsoft said it fixed in December 2020. The original Google PoC works unmodified. May 2026 Patch Tuesday updates do not stop it. The same researcher has dropped several other Windows zero-days in recent weeks, all of which were quickly seen in real attacks.
Microsoft has refused to issue a CVE for what an outside researcher and the CERT Coordination Center both describe as a privilege escalation in Azure Backup for Azure Kubernetes Service. The flaw lets a user holding only the low-privileged 'Backup Contributor' Azure role gain cluster-admin on AKS clusters, which Microsoft dismissed by saying the attacker 'already held administrator access.' CERT/CC validated the bug and tracked it as VU#284781. The researcher says Microsoft also tried to get MITRE to reject the submission as 'AI-generated content,' then quietly added new permission checks, suggesting a silent patch even as Microsoft says 'no product changes were made.'
cPanel released patches Friday for three new vulnerabilities. The two worst (CVE-2026-29202 and CVE-2026-29203, both CVSS 8.8) let authenticated users execute arbitrary Perl code through the create_user API or escalate privileges via unsafe symlink chmod. The third (CVE-2026-29201, CVSS 4.3) lets authenticated users read arbitrary files. No exploitation observed yet. The disclosure lands while attackers are still mass-exploiting CVE-2026-41940 to deploy 'Sorry' ransomware against cPanel hosts, including a wave targeting government agencies and MSPs (covered May 5). Hosting providers face a compounding patch burden.
Researcher Hyunwoo Kim disclosed Dirty Frag yesterday after an unrelated third party broke the embargo five days early. The flaw chains two Linux kernel page-cache write bugs (xfrm-ESP and RxRPC) to give any local user root access on every major distribution - Ubuntu, RHEL, CentOS Stream, AlmaLinux, openSUSE Tumbleweed, Fedora. Like Dirty Pipe and last week's Copy Fail, it's a deterministic logic bug with no race condition required and no kernel panic on failure. PoC is public on GitHub. The ESP variant patch was merged into the netdev tree on May 7 but distribution kernels remain unpatched. No CVE assigned yet because the embargo broke early.
Researchers at Theori and Xint disclosed Copy Fail yesterday, a Linux kernel bug introduced in 2017 that lets any unprivileged user with shell access become root in seconds. The exploit is a 732-byte Python script that works without version-specific tweaks on every major Linux distribution since 2017 - Ubuntu, Amazon Linux, RHEL, SUSE. Unlike previous kernel bugs (Dirty Cow, Dirty Pipe), Copy Fail has no race condition and no per-kernel offsets. It also leaves no trace on disk because it only modifies the in-memory page cache. The bug was found using AI-assisted reverse engineering and has been hiding in the open for nearly nine years.
Microsoft quietly patched a privilege escalation flaw in Entra ID (formerly Azure AD) that let an attacker with a low-privileged service account take over any service principal in the same tenant - including high-value ones with admin consent grants. The bug was in how Entra ID validated role assignments during certain API calls: the validator checked whether the caller had any role on a service principal but didn't check whether that role authorized the specific action. Microsoft fixed the flaw on the back end, so customers don't need a patch - but the takeover scenario means anyone who exploited it before the fix could have created persistent backdoors via OAuth grants.
Kaspersky disclosed PhantomRPC at Black Hat Asia on April 24, an architectural flaw in how Windows handles a core internal communication system called RPC (Remote Procedure Call). When a privileged Windows process tries to talk to an RPC server that isn't running, the operating system doesn't check whether the thing answering is the real one - so a low-privileged attacker can stand up a fake RPC server, intercept the call, and inherit SYSTEM-level access. All Windows versions are affected. Kaspersky demonstrated five different exploitation paths and published the research tools on GitHub. Microsoft has not released a patch.
CISA added CVE-2026-33825 to its Known Exploited Vulnerabilities catalog on April 23 with a May 7 federal patch deadline. The flaw, nicknamed BlueHammer, is a race condition in Windows Defender's file-remediation logic that lets an unprivileged local attacker overwrite arbitrary files on disk and escalate to SYSTEM on fully-patched Windows 10 and Windows 11 hosts. It was patched in Microsoft's April 8 Patch Tuesday but a working proof-of-concept had already been published to GitHub by a researcher called 'Chaotic Eclipse' on April 7, before the fix shipped. Huntress Labs saw in-the-wild exploitation from April 10, with attackers also picking up two sibling Defender zero-days the same researcher leaked: RedSun (another local privilege escalation) and UnDefend (a denial-of-service that blocks Defender from pulling security definition updates, effectively disarming the EDR). Those two still have no Microsoft patch. The combination - a working privilege-escalation path plus an unpatched technique to silently cripple Defender itself - makes this a priority hunt, not just a priority patch.
Microsoft released out-of-band security updates for a critical ASP.NET Core Data Protection flaw that lets unauthenticated attackers forge authentication cookies and escalate to SYSTEM privileges. The bug (CVE-2026-40372) is a regression introduced in the April 2026 Patch Tuesday: the Microsoft.AspNetCore.DataProtection 10.0.0 through 10.0.6 NuGet packages compute the HMAC validation tag (the cryptographic signature that proves a cookie has not been tampered with) over the wrong bytes of the payload and then discard the hash in some cases. The broken check means attackers can forge payloads that pass DataProtection's authenticity checks and decrypt previously-protected data in auth cookies, antiforgery tokens, TempData, and OIDC state. Microsoft noticed the flaw only after users reported decryption failures in their apps after installing the .NET 10.0.6 update. Critical operational detail: updating to 10.0.7 stops future forgeries, but any tokens an attacker already got the app to legitimately sign during the vulnerable window (session refresh tokens, API keys, password reset links) remain valid forever unless you rotate the DataProtection key ring. Patching alone is not enough.
TrustStrike Labs