← Carolina Clear Tech

Cyber Threat Brief

2026-11-14

Listen to this brief (13:45)

Download MP3
Show Notes

Show Notes - 2026-11-14

Stories Covered

CVEs Referenced

CVE-2026-114713, CVE-2026-116322, CVE-2026-123039, CVE-2026-6460

Read the full brief

Get tomorrow's brief in your inbox

Protect Your Business

Need a security assessment? See our cybersecurity packages.

View Services

Today: SCKit is abusing ScreenConnect access to compromise MSP-managed endpoints and target technician sessions; removing unauthorized clients alone is insufficient. CISA added actively exploited UnrealIRCd and Check Point vulnerabilities to KEV, with November 16 remediation dates for covered federal agencies. Prioritize exposed refrigeration controllers and PostgreSQL updates, including rebuilding C-language extensions.

Critical Alerts

CISA adds exploited UnrealIRCd and Check Point flaws to KEV (CVE-2026-114713, CVE-2026-123039)

Summary: CISA confirmed active exploitation of CVE-2026-114713, an UnrealIRCd delimiter-neutralization vulnerability, and CVE-2026-123039, an incorrect-permissions vulnerability affecting multiple Check Point products. KEV enrichment lists November 16, 2026 remediation dates for both; EPSS is 0.3% at the 23rd percentile and 1.3% at the 70th percentile, respectively. Evidence of exploitation takes priority over these comparatively low predictive scores.

SCKit abuses ScreenConnect to compromise endpoints and technician sessions

Summary: Microsoft confirmed an active campaign involving unauthorized access to cloud-hosted ScreenConnect instances belonging to MSPs, followed by unauthorized remote-access clients and the SCKit worm on managed endpoints. SCKit persists through scheduled tasks and registry Run entries; its RelayTrap component targets connecting technicians, while hijacked authenticated sessions enable further distribution. ConnectWise is investigating extension abuse as an entry route, including an RMM+ authentication flaw affecting versions 1.0.31 and 1.0.32 when no access key is configured.

Actively exploited flaws threaten Dixell and Copeland refrigeration controllers

Summary: VulnCheck disclosed 14 vulnerabilities affecting the XWEB EVO supervision system, with lab testing showing root or administrative access through each flaw. CyberScoop says many carry severity scores between 8 and 9.8, and at least two are under active exploitation. Copeland confirmed affected XWEB monitoring solutions and urged customers to remove public internet exposure and install patched software.

Ransomware & Extortion

Operation Endgame disrupts Gootloader, CastleRAT and StealC infrastructure

Summary: Authorities seized 235 servers and 257 domains supporting malware used for credential theft, remote access and ransomware delivery. Germany's BKA said police recovered about 6.4 million stolen logins belonging to more than 40,000 victims; The Hacker News also records substantially different US and German estimates of Gootloader's historical infection count. Infrastructure seizures do not invalidate stolen credentials or establish that affected endpoints are clean.

Business & Infrastructure Threats

ARTEX moves from authorized testing into criminal and espionage operations

Summary: Cisco Talos describes operational misuse of AI-enabled testing frameworks that automate reconnaissance, vulnerability testing and follow-on activity. SentinelOne attributes ARTEX use to financially motivated and state-aligned actors, including attempted sign-ins against more than 2,250 Azure tenants following a ConnectWise cloud-application identity compromise. SentinelOne also describes suspected ARTEX activity moving from a leaked GitHub secret to Azure Key Vault data in five minutes; that observation does not establish the same timeline for every intrusion.

AI-agent containment requires workflow, identity and network controls

Summary: An analysis of agent-response controls explains why disabling a model does not necessarily stop queued jobs, retries, connected tools or existing sessions. It distinguishes suspicious behavior from confirmed compromise and recommends corroborating agent logs with identity-provider, API, application and network evidence. The support-ticket exfiltration scenario is illustrative, not a documented breach.

Windows / AD Security

SignDrop disclosure describes local elevation to SYSTEM through CLFS

Summary: Researcher Nightmare Eclipse disclosed SignDrop, described by Ars Technica as a Windows zero-day granting local users SYSTEM rights on some Windows versions. The attack combines Common Log File System behavior, Defender signature-update processing and manipulated symbolic links to abuse privileged file creation. This is a local privilege-escalation path, not evidence of unauthenticated remote compromise or active exploitation.

Turla's Cereal framework targets Exchange email collection

Summary: SentinelOne describes Turla deploying the Cereal espionage framework against government Exchange environments in Kazakhstan and Romania. Its CerealBox component injects into EdgeTransport.exe and intercepts transport handlers to select, encrypt and stage targeted messages without registering an unauthorized transport agent. The framework uses the compromised Exchange server itself rather than external command-and-control infrastructure, limiting the value of perimeter-only detection.

General Security News

UK healthcare leaders face November 27 safeguard-response deadline

Summary: The ICO wrote to more than 200 senior UK health leaders after assessing over 300 reports of medical-record snooping between January and mid-October 2026. These are organizational reports, not a count of confirmed breaches. Recipients must respond by November 27; leaders without a recent satisfactory safeguard review have another three months to address deficiencies.

Trezor confirms recovery-phrase implants in reseller-supplied wallets

Summary: Trezor identified implants in three Safe 5 wallets purchased from CryptoBilis, examining one in its lab and identifying two through owner photographs. The implants capture displayed recovery information and transmit it over mobile networks; Trezor says firmware 2.13.1 stops the implants found so far from capturing information but cannot protect previously captured backups. The company says none of the three owners has reported lost funds.

Project Zero shortens vulnerability-disclosure timelines

Summary: Project Zero replaced its "90 + 30" policy with disclosure 90 days after vendor notification, removing the separate patch-soak period and most special cases. It will move to 60 days on February 2, 2027, and 30 days on January 3, 2028. Project Zero attributes the change to faster AI-assisted discovery and exploitation; these are its disclosure-policy dates, not universal customer patch deadlines.

Patch Priority

Vulnerability Disclosures

PostgreSQL fixes 52 flaws, including authenticated code execution (CVE-2026-6460, CVE-2026-116322)

Summary: PostgreSQL fixed CVE-2026-6460, a type-confusion vulnerability rated CVSS 8.8 that lets a database user able to create objects execute code with the database service's operating-system privileges. Fixed releases are 18.7, 17.12, 16.16, 15.20 and 14.25; the project has not reported exploitation, and EPSS enrichment lists 0.6%, at the 49th percentile. The release also fixes CVE-2026-116322, an unauthenticated restart vulnerability affecting PostgreSQL 18, with EPSS 0.4%, at the 31st percentile.

Trends & Context

Management access and trusted identities are recurring leverage points: ScreenConnect sessions, cloud-application identities and database object-creation permissions can turn limited access into broader control. Across endpoint malware and agent workflows, containment must remove persistence and invalidate access, not merely stop the visible application. Active exploitation should outrank low EPSS estimates when setting today's remediation queue.