An AI-discovered bug hidden in NGINX since 2008 lets anyone on the internet crash NGINX worker processes or, with ASLR disabled, run code on the server using a single crafted HTTP request. The flaw, named NGINX Rift (CVE-2026-42945, CVSS 9.2), sits in the rewrite module that powers URL rewriting in almost every NGINX deployment. It triggers when a config uses a rewrite directive with unnamed regex captures and a question mark, followed by another rewrite, if, or set directive - a common pattern in API gateway setups. NGINX runs roughly a third of the websites on the public internet.
A researcher who calls themselves Chaotic Eclipse - and who has weaponized every prior Windows flaw they have leaked this year - dropped working proof-of-concept code for two unpatched zero-days on May 12. YellowKey lets anyone with physical access to a Windows 11 or Server 2022/2025 machine plug in a USB stick, hold CTRL during a reboot into the Windows Recovery Environment, and get a shell with full access to the BitLocker-protected drive. GreenPlasma is a privilege escalation against the CTFMON service that hands an unprivileged user a path to SYSTEM. Independent researchers including Will Dormann and Kevin Beaumont have confirmed that YellowKey works as advertised.
Six days after Dirty Frag was patched, researcher William Bowling and the V12 Security team disclosed Fragnesia - a separate Linux kernel bug in the same ESP-in-TCP networking code that lets any unprivileged local user become root in one command. The public proof-of-concept overwrites /usr/bin/su in memory using a logic flaw that loses track of shared socket-buffer fragments, then re-runs su to drop into a root shell. The on-disk binary is left untouched, which makes the change harder to spot. Tracked as CVE-2026-46300 (CVSS 7.8), it follows Copy Fail (April 29) and Dirty Frag (May 7) in the same family.
Quest KACE has a year-old maximum-severity authentication bypass (CVE-2025-32975, CVSS 10.0). Hunt.io researchers now report that an attacker exploited an unpatched KACE appliance at a Boston-area managed services provider called HIQ - then left their entire toolkit on a publicly accessible server with directory listing turned on. The exfiltrated 512 MB MariaDB dump turned out to contain the full appliance-managed endpoint list for over 60 named client organizations spanning law enforcement, government, healthcare, education, and private companies. None of those 60-plus organizations had any KACE relationship of their own - they were just customers of the MSP that ran it unpatched.
Microsoft fixed 120 vulnerabilities on Tuesday - 17 Critical, no zero-days for the first time since June 2024. Two Word RCEs (CVE-2026-40361 and CVE-2026-40364) trigger just by viewing a malicious document in Outlook's Preview Pane and are rated 'Exploitation More Likely.' Windows DNS Client (CVE-2026-41096) lets an attacker-controlled DNS server execute code on any Windows machine resolving a hostile name - echoing SigRed. Other priorities: Netlogon RCE (CVE-2026-41089) and Microsoft SSO Plugin for Jira and Confluence (CVE-2026-41103, CVSS 9.1).
Fortinet patched two critical RCE flaws Tuesday. CVE-2026-44277 in FortiAuthenticator (Fortinet's IAM/MFA platform) lets unauthenticated attackers execute code via crafted requests. CVE-2026-26083 (CVSS 9.1) in FortiSandbox's web UI lets unauthenticated attackers run code via HTTP requests. Neither is confirmed exploited yet, but Fortinet products have a long exploitation history - CISA flagged FortiClient EMS as actively exploited in April. FortiSandbox is the threat-detection backbone for many Fortinet-centric SOCs; FortiAuthenticator gates MFA and SSO.
SAP's May Patch Day included two CVSS 9.6 critical flaws. CVE-2026-34263 in Commerce Cloud is a missing authentication check from improperly ordered Spring Security rules - unauthenticated attackers can upload configurations and inject code. CVE-2026-34260 in S/4HANA is a SQL injection in the ABAP Enterprise Search component that lets low-privilege authenticated users steal sensitive database records. Both land less than two weeks after four SAP npm packages were hit in the Mini Shai-Hulud attack, putting SAP customers under compounding patch pressure.
Google's Threat Intelligence Group says it caught the first known case of a real attacker using a large language model to find and weaponize a zero-day - a 2FA bypass in a popular but unnamed open-source web-based system administration tool. Google has high confidence the Python exploit was AI-generated, citing textbook code structure, abundant educational docstrings, and a hallucinated CVSS score in the script. The flaw was a high-level logic bug, the kind LLMs excel at spotting, rather than a memory corruption issue. Google rules out Gemini and warns that AI-assisted exploit development is being industrialized via account-pooling and proxy relays for premium models.
Researchers at Cyera disclosed a critical bug in Ollama, the open-source tool that runs large language models locally on laptops and servers. The flaw, called Bleeding Llama (CVE-2026-7482), lets anyone with network access send a malformed model file and read raw process memory back - which typically contains API keys, environment variables, system prompts, and other users' chat history. Ollama ships without authentication by default, so an estimated 300,000 instances are exposed on the internet. Ollama 0.17.1 fixes it. Separately, Striga disclosed two unpatched Ollama Windows desktop flaws (CVE-2026-42248 and CVE-2026-42249) that chain into persistent code execution at login.
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.
TrustStrike Labs