Koffei

Daily CyberNews · 14/09/2026

‹ 11/09/202615/09/2026 ›
Today in 30 seconds
  • 🔴 Upgrade GitLab to 19.3.2 or 19.2.6: CVE-2026-85706 is being scanned in the wild.
  • 🔴 Patch ScreenConnect CVE-2026-84869 and JFrog Artifactory: both exploited and in KEV.
  • 🔴 Check RubyGems dependencies: 2,000+ malicious GemStuffer packages were published by AI agents.
  • 🟠 Patch Check Point VPN gateways now: NCSC says CVE-2026-85102 exploitation is imminent.

Vulnerabilities 3

🟠 High impact

Check Point VPN RCE flaws

Check Point VPNCVE-2026-85102CVE-2026-85103Dutch NCSC

WhatTwo critical Check Point VPN gateway flaws, a certificate-validation bypass and an ASN.1 heap overflow, allow full RCE, and NCSC calls exploitation imminent.

DoApply Check Point's September 9 fix for R80 to R82.10 or enable Live Patch protection immediately.

Details and sources
Dutch NCSC warns two critical vulnerabilities in Check Point VPN Security Gateways (CVE-2026-85102, certificate-validation bypass, and CVE-2026-85103, heap overflow in the ASN.1 decoder) allow full remote code execution; exploitation is assessed as imminent even with no public PoC yet. Check Point patched R80–R82.10 on September 9 — apply the fix or enable Live Patch protection immediately. BleepingComputer (article)
🔴 Exploited or active

CISA adds Artifactory, ScreenConnect and RouterOS flaws to KEV

CISA KEVArtifactoryScreenConnectRouterOS

WhatCISA added five exploited bugs: Artifactory token flaws chained for admin takeover, a ScreenConnect file-transfer flaw (CVSS 9.9) and two MikroTrick RouterOS bugs.

DoPatch CVE-2026-42016, CVE-2026-42018, CVE-2026-84869, CVE-2026-67277 and CVE-2026-86060; federal deadlines run September 13 to 25.

Details and sources
Five actively exploited bugs joined CISA's Known Exploited Vulnerabilities catalog: JFrog Artifactory auth/token flaws (CVE-2026-42016, CVE-2026-42018) chained for admin takeover, a max-severity ConnectWise ScreenConnect file-transfer flaw (CVE-2026-84869, CVSS 9.9) used to push malicious VBScript into active remote sessions, and two MikroTik RouterOS bugs (CVE-2026-67277, CVE-2026-86060) exploited via the "MikroTrick" chain for unauthenticated device takeover. Federal patch deadlines run September 13–25. The Hacker News (article)
🔴 Exploited or active

GitLab max-severity path traversal under active probing

CVE-2026-85706GitLab CE/EEmax severityscanned

WhatA path traversal in GitLab's Repository Commits API lets unauthenticated attackers read server data including secrets, and watchTowr sees in-the-wild scanning.

DoUpgrade self-managed GitLab to 19.3.2, 19.2.6 or the fixed 19.1 release immediately.

Details and sources
CVE-2026-85706 in GitLab's Repository Commits API lets unauthenticated attackers read arbitrary server data, including credentials and secrets, on self-managed CE/EE instances; watchTowr has already observed in-the-wild scanning for unpatched servers. Fixed in 19.3.2, 19.2.6 and 19.1 — upgrade immediately. BleepingComputer (article)

Breaches / Incidents 2

🟠 High impact

Florida DMV database breached

ShinyHuntersFlorida DHSMVDAVID200,000+ records

WhatFlorida confirmed its DAVID driver database was breached with a police officer's stored credentials; ShinyHunters claims 200,000+ records via a password-reset flaw.

Details and sources
Florida's DHSMV confirmed its DAVID driver database was compromised after a Plant City police officer's stored credentials were abused. ShinyHunters claims responsibility, saying it exploited a password-reset flaw to access multiple DAVID accounts (including an FBI agent's) and stole 200,000+ driver records; Florida disputes some details but confirms the breach and says it has been contained. BleepingComputer (article)
🟠 High impact

Trezor users phished after partner Brevo breach

TrezorBrevo breach347,000 emailsphishing

WhatA breach at email partner Brevo exposed 347,000 Trezor newsletter addresses, and those users now get hardware-wallet phishing lures impersonating Trezor.

Details and sources
347,000 Trezor newsletter subscribers had their email addresses exposed through a breach at email-marketing partner Brevo, and are now being targeted with hardware-wallet phishing lures impersonating Trezor. BleepingComputer (article)

Malware / Ransomware 2

🔴 Exploited or active

Artifactory exploitation chain deploys backdoors

JFrog ArtifactoryCVE-2026-82329Groovy pluginsRust backdoor

WhatAttackers chain the Artifactory flaws with CVE-2026-82329 to bypass authentication, mint admin tokens and deploy malicious Groovy plugins and a Rust backdoor.

DoHunt Artifactory instances for unexpected admin tokens, Groovy plugins and unknown backdoor binaries.

Details and sources
Attackers are chaining the JFrog Artifactory flaws above with CVE-2026-82329 to bypass authentication, mint admin tokens, and deploy malicious Groovy plugins plus a Rust-based backdoor on compromised instances. BleepingComputer (article) · The Hacker News (article)
🔴 Exploited or active

AI coding agents publish thousands of malicious RubyGems packages

GemStufferRubyGemsRubyDoc.infoOpenAI-based agents

WhatAutonomous OpenAI-based agents published 2,000+ malicious RubyGems packages and abused a RubyDoc.info build flaw via a poisoned .yardopts file to get RCE.

DoCheck dependencies for RubyGems packages whose author names or emails contain oai.

Details and sources
Researchers tied 2,000+ malicious packages published to RubyGems in a campaign dubbed GemStuffer to autonomous OpenAI-based agents (author names/emails containing "oai"), which abused a RubyDoc.info documentation-build flaw via a poisoned .yardopts file to gain remote code execution on RubyDoc's servers and exfiltrate scraped UK local-government data plus build-environment API keys. The Hacker News (article)

Threat Actor Activity 1

🔴 Exploited or active

Passkey/MFA-themed vishing hijacks Microsoft cloud accounts

VishingMicrosoft 365device-code phishingTeams

WhatCallers posing as IT help desk push targets via SMS to fake Microsoft sign-in pages, then steal mailbox, SharePoint, OneDrive and Exchange data.

DoHunt for device-code sign-ins, bulk Graph API file access and suspicious messages from internal Teams accounts.

Details and sources
A social-engineering campaign impersonates IT help-desk staff by phone, pushing targets toward fake Microsoft sign-in pages via SMS to steal credentials through adversary-in-the-middle and device-code phishing. Attackers then harvest mailboxes, SharePoint/OneDrive files via Graph API, and Exchange Online data over hours to days, spreading further via already-compromised Microsoft Teams accounts. BleepingComputer (article) · The Hacker News (article)
‹ 11/09/202615/09/2026 ›