This website uses cookies

Read our Privacy policy and Terms of use for more information.

Attackers are exploiting an unauthenticated code execution flaw in Confluence Data Center when its Remote API is enabled; Atlassian lists version 9.2.28 as fixed. The attacks put Confluence ahead of other major patching concerns: NoisyTCP can crash Linux servers in default configurations, while an Apache HTTP/2 flaw has a demonstrated remote code execution exploit but no confirmed attacks.

The Linux and Apache flaws expose widely used server software without requiring attackers to log in, though their consequences differ: NoisyTCP causes a kernel panic, not code execution. Fixes are available for both through Linux kernel updates and Apache HTTP Server 2.4.71, but affected standalone mod_h2 releases have no standalone fix.

Other attacks relied on familiar names and trusted services: phishing emails claimed Microsoft 365 messages faced deletion, while 186 lookalike domains impersonated AI brands to distribute malware. Infoblox also confirmed that attackers accessed company information through a compromised data migration vendor, rather than through a confirmed breach of its products or cloud services.

An 11-year-old Linux kernel flaw, NoisyTCP (CVE-2026-99288), affects almost all Linux distributions and lets unauthenticated attackers crash servers running Nginx, Apache or SSH in default configurations using crafted TCP packets. The flaw causes a kernel panic but does not allow data modification or arbitrary code execution. A patch is available, and the fix entered the mainline kernel two weeks ago.

A race condition in Apache HTTP Server’s mod_http2 module, CVE-2026-119618, allows unauthenticated remote code execution in versions 2.4.55–2.4.70 with HTTP/2 enabled. Apache 2.4.71 fixes the flaw; standalone mod_h2 versions 2.0.8–2.0.42 are also affected and have no standalone fix. Public research demonstrated a reliable exploit, but no confirmed exploitation in the wild has been reported.

Microsoft detected a high-volume phishing campaign on November 13 targeting organizations worldwide with fake Microsoft 365 outage notifications. The emails claimed four messages would be deleted within 24 hours unless recipients clicked a recovery link; multiple samples led to reallifeaccounting[.]com. Microsoft did not confirm credential theft or account compromise and has not attributed the campaign to a named threat actor.

Attackers used a shipment-themed ZIP archive to load a malicious DLL through a legitimate, signed Kingsoft PDF application and deploy a Windows payload. Unit 42 observed the activity on October 21, 2026; the chain used disguised WinRAR and Python utilities and encoded code hidden in a fake PDF. The payload contacted infrastructure associated with PureLogs and PureRAT campaigns.

Attackers are exploiting CVE-2026-90436 to execute code without authentication in Atlassian Confluence Data Center versions 1.x through 9.x when the Remote API is enabled. Rapid7 detected exploitation on November 13; Atlassian lists version 9.2.28 as fixed. Censys counted 1.2 million internet-facing Confluence hostname-and-port combinations, not confirmed vulnerable installations or separate servers.

Mozilla released Firefox 160 to fix 99 security vulnerabilities, including 57 rated high impact and multiple sandbox escapes. The fixes include CVE-2026-124907, a use-after-free flaw in CanvasWebGL that can allow privilege escalation and sandbox escape, and CVE-2026-124903, a sandbox escape in interprocess communication. Mozilla’s advisory does not report active exploitation.

Unicode filename handling flaws let local attackers bypass macOS file exclusion and permission checks in Deno, libarchive, HashiCorp go-slug and Python setuptools, potentially overwriting files or exposing secrets in packages. Four CVEs cover the flaws, including CVE-2026-59890 in setuptools versions before 83.0.0, which fixes the issue. CERT/CC reports no exploitation in the wild.

Attackers used 186 typo-squatted domains impersonating 27 AI brands, including ChatGPT, Claude and Perplexity, to distribute malware to users who mistype web addresses. The sites used ClickFix prompts to induce users to run harmful PowerShell commands, fake security checks to deliver a browser hijacker, and malicious extensions that manipulate search results and collect user data. Many spoofed domains remain active.

Hackers accessed Infoblox company information through a compromised third-party data migration vendor. TechCrunch reported that the vendor provided SharePoint migration services and that sensitive customer data was taken, citing a person familiar with the incident. Infoblox did not disclose how many customers were affected and said it had no indication its products, cloud services or software release processes were involved.

Apple patched CVE-2026-65339, a flaw that let a malicious sandboxed app record microphone audio despite a user denying permission. The app could impersonate Apple’s audio daemon to bypass consent checks; on an earlier iOS 27 beta, recording also left the microphone indicator off. Public fixes first shipped August 17 in iOS 26.6.1, iPadOS 26.6.1 and macOS Tahoe 26.6.2.

Quick Hits

Reply

Avatar

or to participate