Koffei

Daily CyberNews · 07/10/2026

‹ 06/10/202608/10/2026 ›
Today in 30 seconds
  • 🔴 Patch Ninja Forms to 3.15.4 and WPC Product Bundles to 8.6.7, then check sites for backdoors.
  • 🔴 Rejetto HFS CVE-2026-61500 is being actively scanned; upgrade to 3.2.1.
  • 🔴 LunexStealer spreads via fake Cloudflare checks on 100+ sites; watch for the fake Word Editor extension.
  • 🟠 Patch Atlassian Data Center for CVE-2026-21589 and on-prem Exchange for CVE-2026-96940.
  • 🟠 Update LibreOffice to 26.2.5.2-1; avoid untrusted spreadsheets in OpenOffice until 4.1.17.

Vulnerabilities 6

🟠 High impact

Atlassian Data Center critical file-access flaw

CVE-2026-21589CVSS 9.3Atlassian Data Center

WhatAn unauthenticated attacker can read arbitrary known files across eight self-hosted Atlassian Data Center products; EPSS is very low and it is not in KEV.

DoInstall patched releases, for example Confluence Data Center 9.2.26/10.2.19 or Jira Software 9.12.40/10.3.26/11.3.12.

Details and sources
CVE-2026-21589 (CVSS 9.3 on the CVSS 4.0 vector per NVD/OpenCVE — CWE-552/CWE-200; cyberveille.esante.gouv.fr separately lists it at CVSS 7.5) lets an unauthenticated attacker read arbitrary known files across eight self-hosted Atlassian Data Center products: Bitbucket, Confluence, Crowd, Jira Software, Jira Service Management, Bamboo, Crucible and Fisheye. EPSS is very low (<1%) and it is not yet in the CISA KEV catalog, but patched releases are available for every affected product (e.g. Confluence Data Center 9.2.26/10.2.19, Jira Software 9.12.40/10.3.26/11.3.12). BleepingComputer (source) · The Hacker News (source)
🟠 High impact

Microsoft Exchange mailbox-read flaw

CVE-2026-96940CVSS 8.8Exchange on-prem

WhatAn authenticated attacker can elevate privileges and read other users' mailboxes on on-prem Exchange 2016 CU23, 2019 CU14/CU15 and Subscription Edition RTM.

DoInstall Microsoft's fixed Exchange builds for every affected branch.

Details and sources
CVE-2026-96940 (CVSS 8.8, CWE-1390 weak authentication) lets an already-authenticated attacker elevate privileges and read other users' mailboxes on on-premises Exchange Server 2016 CU23, Exchange 2019 CU14/CU15, and Exchange Server Subscription Edition RTM. EPSS is very low and it is not in CISA KEV; Microsoft has shipped fixed builds for every affected branch. The Hacker News (source)
🔴 Exploited or active

Rejetto HFS critical RCE now being actively scanned for

CVE-2026-61500CVSS 9.8Rejetto HFSactive scanning

WhatAttackers are scanning for HFS servers where the cookie signing key can be derived from Math.random() outputs to forge admin sessions and get RCE.

DoUpgrade Rejetto HFS 3.0.0 to 3.2.0 to version 3.2.1.

Details and sources
Threat actors are scanning the internet for CVE-2026-61500 (CVSS 9.8 on v3.1 / 9.3 on v4.0 — CWE-338, weak PRNG) in Rejetto HFS 3.0.0–3.2.0. The flaw lets a remote attacker derive the session-cookie signing key from exposed Math.random() outputs, forge a valid administrator session, and reach full remote code execution via the server_code configuration feature. Not yet in CISA KEV despite the active scanning; fixed in HFS 3.2.1. BleepingComputer (source)
🟡 Worth knowing

LibreOffice / OpenOffice spreadsheets can run code without a macro warning

CVE-2026-63277CVE-2026-59265LibreOfficeOpenOffice

WhatMalicious spreadsheets can run code on open without a macro warning in LibreOffice and Apache OpenOffice, with OpenOffice still effectively unpatched.

DoUpdate LibreOffice to 26.2.5.2-1 and treat untrusted spreadsheets in OpenOffice 4.1.16 or earlier with caution until 4.1.17.

Details and sources
Two office-suite flaws let a malicious spreadsheet execute code on open without the usual macro-security prompt. CVE-2026-63277 affects LibreOffice 7.4.5-3 through 26.2.5.1-2 and is already fixed in 26.2.5.2-1 (tracked at roughly CVSS 4.0 by vendor trackers). CVE-2026-59265 affects Apache OpenOffice 4.1.16 and earlier (CWE-426, untrusted search path; no CVSS has been assigned yet) and remains effectively unpatched — a fix is only at release-candidate stage (4.1.17 RC). Neither is in CISA KEV. The Hacker News (source)
🔴 Exploited or active

WordPress plugin XSS flaws exploited to plant backdoors

CVE-2026-94504CVE-2026-93836Ninja FormsWooCommerce

WhatAttackers exploit stored XSS in Ninja Forms and WPC Product Bundles to create hidden admin accounts and unauthenticated file managers.

DoUpdate Ninja Forms to 3.15.4 and WPC Product Bundles to 8.6.7, then manually check for rogue admins and backdoors.

Details and sources
Attackers are exploiting stored XSS flaws in two popular WordPress plugins to create hidden admin accounts and unauthenticated file managers. CVE-2026-94504 (Ninja Forms, CVSS 7.2, CWE-79) affects versions through 3.15.3 (500,000+ sites), fixed in 3.15.4. CVE-2026-93836 (WPC Product Bundles for WooCommerce, CVSS 7.2, CWE-79) affects versions through 8.6.6 (30,000+ sites), fixed in 8.6.7. Neither is in CISA KEV, and patching alone does not remove existing backdoors — site owners must manually check for compromise. BleepingComputer (source)
⚪ Context

Pwn2Own Ireland — 32 zero-days on day one

Pwn2Own Ireland32 zero-daysGalaxy S26

WhatResearchers earned $388,500 on day one of Pwn2Own Ireland, demonstrating 32 zero-days including two Samsung Galaxy S26 compromises.

Details and sources
Researchers earned $388,500 after compromising a Samsung Galaxy S26 twice among 32 zero-day exploits demonstrated on the first day of the contest. BleepingComputer (source)

Breaches / Incidents 2

🟡 Worth knowing

ASOS confirms data breach after hijacked app notifications

ASOSXuanye grouppush notificationsSnowflake

WhatAttackers abused ASOS's third-party notification platform to send fake alerts and claim Snowflake data theft; ASOS confirmed names and contact details may be exposed.

Details and sources
Attackers calling themselves the "Xuanye group" abused ASOS's third-party notification platform to blast unauthorized push alerts ("ASOS HACKED") to customers, and separately claimed to have stolen data from the retailer's Snowflake environment. ASOS confirmed names and contact details may have been exposed but says payment-card data and account passwords were not compromised; it has not confirmed the Snowflake claim. BleepingComputer (source)
🟡 Worth knowing

Nikkei discloses employee email breaches

Nikkeiemployee email accountsphishing

WhatAttackers breached two Nikkei employees' Microsoft and Google email accounts and used at least one to send phishing messages.

Details and sources
Japanese publisher Nikkei confirmed unknown attackers breached two employees' Microsoft and Google email accounts and used at least one of them to send phishing messages. BleepingComputer (source)

Malware / Ransomware 6

🔴 Exploited or active

LunexStealer spreads via fake Cloudflare checks

LunexStealerLUNARAXENAIVEMESSUAC-0277

WhatOver 100 compromised sites use fake Cloudflare checks to install LunexStealer via a malicious Microsoft Office Word Editor browser extension, with one variant abusing an AMD driver for UAC bypass.

DoHunt for the fake Microsoft Office Word Editor extension and block ClickFix-style Cloudflare verification lures.

Details and sources
Over 100 compromised websites are using the ClickFix technique — forged Cloudflare "verify you're human" pages — to trick Windows users arriving from search engines into installing LunexStealer through a malicious browser extension, LUNARAXE, that masquerades as a "Microsoft Office Word Editor" add-on, alongside a file-access component, NAIVEMESS. One MSI variant bypasses Windows UAC via a vulnerable AMD driver. CERT-UA attributes the campaign, observed in September 2026, to the threat cluster UAC-0277. The Hacker News (source)
🟠 High impact

Fake AI ad portals steal credentials and MFA codes

Fake AI ad portalsChatGPTGeminiClaudeBitB

WhatA phishing platform imitating ChatGPT, Gemini and Claude ad portals uses browser-in-the-browser pages to capture credentials and MFA codes.

Details and sources
A phishing platform impersonating ChatGPT, Gemini, and Claude advertising/business portals uses a browser-in-the-browser technique to capture login credentials and MFA codes from victims. The Hacker News (source)
🟠 High impact

Linux backdoors impersonate email security tools

Linux backdoorstelecomSouth KoreaTaiwan

WhatNew Linux backdoors targeting telecom and network appliances in South Korea and Taiwan disguise C2 traffic as email-security services.

Details and sources
New Linux backdoors targeting telecom and network appliances in South Korea and Taiwan disguise their command-and-control traffic as legitimate email-security services to evade detection. The Hacker News (source)
🔴 Exploited or active

ClingSTUN turns vulnerable IoT devices into proxy nodes

ClingSTUNIoTSTUN C224 vulnerabilities

WhatThe ClingSTUN Linux backdoor exploits 24 known flaws in D-Link, Realtek, Ivanti and TP-Link devices and hides C2 traffic using public STUN servers.

DoPatch exposed routers, cameras and industrial devices, and watch for unusual STUN traffic from them.

Details and sources
FortiGuard Labs named ClingSTUN, a Linux backdoor exploiting 24 known vulnerabilities in routers, network equipment, surveillance systems, and industrial devices from D-Link, Realtek, Ivanti, and TP-Link, plus 7 additional hardcoded exploits for lateral spread. It abuses legitimate public STUN servers to map external networks and hide its C2 traffic among normal VoIP/WebRTC traffic. Dark Reading (source)
🟡 Worth knowing

ClickFix attacks evolve to better hide payloads

ClickFixobfuscation

WhatClickFix fake verification prompts are adding new obfuscation layers to hide their malicious clipboard payloads.

Details and sources
Researchers report the ClickFix social-engineering technique — fake verification prompts that trick users into running malicious clipboard commands — is adding new obfuscation layers to better conceal its payloads from defenders. Dark Reading (source)
🟡 Worth knowing

"BigDiskBuster" leaves Defender running while blocking its updates

BigDiskBusterMicrosoft Defenderupdate blocking

WhatA newly documented technique keeps Defender looking active while silently blocking signature updates.

DoMonitor Defender signature update age across endpoints to spot blocked updates.

Details and sources
A newly documented technique keeps Microsoft Defender appearing active while silently preventing it from downloading new signature updates, giving malware a stealthy window to operate undetected. Dark Reading (source)

Threat Actor Activity 2

🔴 Exploited or active

China-aligned TA419 impersonates US officials for AI-policy espionage

TA419ChinaAI policyFrameless BitB

WhatTA419 impersonated a former OSTP deputy director and an economist to lure US policy groups into AitM OneDrive pages that hijack MFA-protected sessions.

Details and sources
Proofpoint reports TA419 ran a months-long social-engineering campaign impersonating a former White House OSTP deputy director and a prominent economist to lure US think tanks, universities, law firms, and defense contractors into a fake "AI Policy Advisory Committee." The group used adversary-in-the-middle phishing with forged OneDrive pages and a custom browser-in-the-browser tool ("Frameless BitB") to hijack MFA-protected sessions, aiming to gather intelligence on US AI-policy development amid US-China strategic competition. Dark Reading (source)
🟡 Worth knowing

Attackers keep abusing legitimate RMM tools for initial access

SANS ISCAction1 RMMVBS phishing

WhatPhishing with fake PDF invoices and DHL lures delivers VBS scripts that install the legitimate Action1 RMM tool for remote access.

DoAlert on unexpected Action1 or ScreenConnect installs and block unapproved RMM tools.

Details and sources
SANS ISC documents phishing campaigns — fake PDF invoices and a DHL-themed variant — that deliver malicious VBS scripts which install the legitimate Action1 remote-monitoring-and-management tool, echoing earlier abuse of ScreenConnect to gain remote access via free or trial accounts. SANS ISC (source)
‹ 06/10/202608/10/2026 ›