Saturday, November 14, 2026 · 1 MIN READ
Active Exploits, Hidden Implants, and Trust Under Attack
IT BRIEFING FOR BUSINESSES
Today’s news puts patch prioritization and access control at the top of the business IT agenda. CISA reports active exploitation of flaws affecting UnrealIRCd and Check Point products, while PostgreSQL has released fixes for 52 vulnerabilities, including one that can let a database user execute code on the server. Ask your IT provider to check whether these products support your networks or business applications and prioritize the relevant updates. For PostgreSQL, confirm that installed C-language extensions are rebuilt where required for full protection. Before maintenance, verify that backups are recoverable and that someone owns the recovery plan—not just the update schedule.
The other stories challenge trust in devices, communications, and automation. Reported attacks include deepfake video calls used in cryptocurrency theft, modified hardware wallets that transmit recovery phrases, and compromised telecom equipment used to intercept calls and texts. Train staff to independently verify sensitive requests rather than relying on a familiar face or message, and review authentication choices for email and Microsoft 365 rather than treating SMS codes as sufficient protection. For AI tools connected to business records, limit permissions and establish how to revoke access and stop actions. Businesses using confidential cloud computing should also ask providers about the newly disclosed AMD SEV-SNP attacks and their requirements for hypervisor and BIOS control.
CISA Flags Two Actively Exploited Vulnerabilities
CISA added CVE-2026-114713 in UnrealIRCd and CVE-2026-123039 in multiple Check Point products to its Known Exploited Vulnerabilities catalog following evidence of exploitation. The UnrealIRCd flaw enables server-level command injection and carries a severity score of 8.6; the Check Point issue involves incorrect permissions on a critical resource.
SOURCES:CISA Advisories The Hacker News
PostgreSQL Fixes 52 Security Flaws
PostgreSQL released updates addressing 52 security flaws, including a 22-year-old type confusion vulnerability tracked as CVE-2026-6460 with a severity score of 8.8. The flaw allows a database user with object-creation privileges to execute code on the server. Full protection also requires rebuilding installed C-language extensions.
SOURCES:The Hacker News
Two Attacks Target Protected Cloud VMs
Researchers disclosed Mior and Skyfall, two attacks that allow a malicious host to read data inside AMD SEV-SNP-protected virtual machines. Mior can additionally alter VM memory and falsify the signed report identifying its software. Both attacks require hypervisor control, while Mior also requires BIOS control.
SOURCES:The Hacker News
Crypto Theft Uses Deepfakes and Wallet Implants
The FBI attributed more than $400 million in cryptocurrency theft over 18 months to the FaceStealer social engineering scheme, which uses deepfake video calls. Separately, Trezor confirmed that three Safe 5 wallets purchased through reseller CryptoBilis contained extra circuit boards designed to capture recovery phrases from the screen and transmit them over a mobile network to an attacker-controlled server.
SOURCES:The Hacker News The Hacker News
AI Agent Access Raises Containment Questions
An analysis examined how organizations can stop an AI agent whose authorized tools and identity permit unwanted actions. Its illustrative scenario describes an agent processing a support ticket containing instructions to export account records to an unfamiliar destination, raising questions about containment when existing permissions allow the request.
SOURCES:The Hacker News
SilentAir Intercepts Calls and SMS Codes
Enea identified a spying group it named SilentAir that used compromised mobile-operator equipment to intercept selected individuals’ calls and text messages, including SMS two-factor authentication codes. The company assessed with high confidence that the group is a state-level actor.
SOURCES:The Hacker News