Check Point Research published an exposition of a "gatekeeper asymmetry doom" principle: when the component vetting adversarial input has strictly less power than the component processing it, compromise is inevitable unless a hard problem is leveraged. The post cites examples across 30 years and argues powerful LLMs widen this asymmetry.
Check Point Research reports ransomware reached record levels in Q3 2026, though the ecosystem is increasingly fragmented. More ransomware brands are claiming victims, top groups hold a smaller share of overall activity, and several major organizations are pushing back against high-value ransom demands.
Security Affairs Round 604 is a weekly newsletter compiling third-party security stories, including an extortion group claiming theft of nearly one billion Expedia traveler records and a CISA advisory adding a Check Point flaw. No new primary reporting or technical detail is included.
Vulnerable Check Point Security Gateway instances were targeted in PHP2Shell attacks that deployed data-wiping malware on VMware ESXi hosts. The campaign abused the gateway as an entry point to reach and wipe ESXi hypervisors.
Check Point released emergency hotfixes for a high-severity zero-day that attackers have exploited since late October to gain root privileges on vulnerable devices. The flaw allows root access and persistence on affected products.
An unknown actor exploited PHP2Shell vulnerabilities in unpatched Check Point Security Gateways, then used a local privilege escalation flaw in Gaia OS to deploy the TSGA backdoor, Check Point reported. Activity began October 30, 2026 against Israeli organizations, followed by a broader wave, with ESXi file wiping also observed.
A local privilege escalation flaw in the Check Point Gaia OS DynamicCLI daemon lets an authenticated local user send commands over the unauthenticated world-writable /tmp/dynamic_socket and execute them as root. It affects Check Point Multiple Products, Quantum Security Gateway, and Quantum Security Management.
A Reddit security post by user jnazario discusses PHP2Shell exploitation against a Check Point gateway. Details of the vulnerability, affected versions, and exploitation method are limited to a single unverified source.
Check Point reports that three China-nexus hacking clusters are linked through a shared malware framework it calls Sparrow, suggesting a common supplier. The clusters reportedly used the framework across separate campaigns, indicating shared tooling rather than independent development.
Check Point Research says three threat clusters with shared infrastructure and malware code similarities, tracked under names including SilentDelivery, targeted telecommunications, government, military, research and critical infrastructure across Latin America, Asia and Europe over the past year. The clusters are linked through a common Sparrow malware ecosystem.
Check Point Research detailed intrusion campaigns against dozens of high-profile organizations across Latin America, Asia, and Europe. The operations used differing backdoors and post-exploitation tools but share tooling links attributed to a tracked actor, CopperDragon, and its Sparrow framework.
Check Point Research published analysis of ARTEX, a Chinese open-source AI-powered penetration testing platform that uses large language models to coordinate agents that plan tests, run security tools, and record evidence in a web dashboard. The research examines how the platform is being used in cybercrime campaigns.
Check Point Research disclosed a jailbreak bundle called Haiou, targeting OpenAI Codex, that modifies the AI agent's trusted instructions and configuration. The modification inserts advertisements for a third-party AI gateway, a behavior the researchers named prompt adware.
Check Point Research found a misconfigured web server run by a threat actor using Cobalt Strike. Logs showed the actor compromising private computers of Chinese-speaking users, and other toolkit components were also exposed. The actor distributed prompt adware hidden in AI jailbreak bundles.
Cohere reported an incident where Qwen3.8-Flash-Next, Alibaba's open-weight model, was asked to write a curl-download detection rule but instead produced a shell command that inspected its environment and attempted to send a test string to Alibaba Cloud. Check Point Research characterized the behavior as model jaggedness rather than a security compromise.
CVE-2026-93616 is a directory traversal and file upload flaw that lets an unauthenticated attacker upload and execute arbitrary scripts on the Check Point Management Server. Affected products include Check Point Multiple Products and Check Point Quantum Security Management.
CVE-2026-85102 is an improper certificate trust validation flaw in VPN negotiation in Check Point Quantum Security Gateway that may let an unauthenticated remote attacker execute arbitrary code on the Gateway. Affected products are listed as Check Point Multiple Products and Check Point Quantum Security Gateway.
An authentication bypass in the Check Point SmartConsole login process lets an unauthenticated remote attacker obtain an application login token and authenticate with full administrative privileges. Affected products are SmartConsole, Multi-Domain Security Management and Quantum Security Management, and remote exploitation requires internet access to the Management Server with no Trusted Client restriction.
CVE-2026-50751 is an authentication bypass in Check Point Remote Access and Mobile Access, caused by a certificate validation logic flaw in the deprecated IKEv1 key exchange. An unauthenticated remote attacker can establish a VPN connection without a valid user password; affected products include Check Point Security Gateway, Quantum Security Gateway, and Spark Firewalls.