OpenAI’s ChatGPT Lockdown Mode is a useful reminder that prompt-injection defense is not just about model behavior. It is about limiting outbound paths, connector permissions, and tool access around sensitive work.
Unit 42 reports active exploitation attempts against PAN-OS GlobalProtect CVE-2026-0257. Defenders should patch, but also review VPN sessions, authentication override cookie behavior, and edge-device telemetry for signs of unauthorized access.
Mandiant reports that UNC3753, also known as Luna Moth / Silent Ransom Group, is targeting U.S. law firms and professional services with vishing, RMM abuse, rapid data theft, and suspected physical office intrusions. Here is what SMBs and government contractors should lock down now.
Cisco says CVE-2026-20245 has been exploited against Catalyst SD-WAN Manager. Defenders should preserve evidence, review controller logs, validate edge-device configuration, and restrict management-plane access.
OpenAI’s ChatGPT Lockdown Mode is a useful reminder that prompt-injection defense is not just about model behavior. It is about limiting outbound paths, connector permissions, and tool access around sensitive work.
Unit 42 reports active exploitation attempts against PAN-OS GlobalProtect CVE-2026-0257. Defenders should patch, but also review VPN sessions, authentication override cookie behavior, and edge-device telemetry for signs of unauthorized access.
Mandiant reports that UNC3753, also known as Luna Moth / Silent Ransom Group, is targeting U.S. law firms and professional services with vishing, RMM abuse, rapid data theft, and suspected physical office intrusions. Here is what SMBs and government contractors should lock down now.
Cisco says CVE-2026-20245 has been exploited against Catalyst SD-WAN Manager. Defenders should preserve evidence, review controller logs, validate edge-device configuration, and restrict management-plane access.
Microsoft’s updated agentic AI failure-mode taxonomy turns AI agents into a practical security architecture problem: plugins, prompts, memory, browser use, and human approvals all need controls.
Group-IB documented a global smishing operation using fake error pages, geofencing, and encrypted WebSocket exfiltration. Here is what SMBs and government contractors should take from it.
Federal agencies warn that attackers are compromising internet-exposed automatic tank gauge systems. The lesson for SMBs, fuel operators, farms, logistics firms, and gov contractors is simple: small OT is still operational infrastructure.
A five-month espionage campaign against a stock exchange executive mailbox shows why senior email accounts need privileged-asset controls, cloud exfiltration monitoring, and scheduled-task hunting.
Proofpoint’s TA4922 reporting shows how localized HR, payroll, tax, and invoice lures can become full initial-access infrastructure through DLL sideloading, loaders, RATs, RMM tools, and browser credential theft.
A Red Hat Cloud Services npm compromise shows why signed releases and trusted publishing must be paired with install-time controls, CI/CD isolation, and fast credential rotation.
