Edge devices that are no longer supported have been targeted in attacks by state-sponsored hackers, the US says. The post Organizations Urged to Replace Discontinued Edge Devices appeared first on SecurityWeek .
Living off the AI isn’t a hypothetical but a natural continuation of the tradecraft we’ve all been defending against, now mapped onto assistants, agents, and MCP. The post Living off the AI: The Next Evolution of Attacker Tradecraft appeared first on SecurityWeek .
The Cyberspy Group is an advanced persistent threat (APT) actor that has conducted cyber espionage operations targeting governments and critical infrastructure across 37 countries. Although Palo Alto Networks has not officially attributed this group, evidence suggests a likely connection to China. The campaign involves sophisticated intrusion techniques aimed at compromising sensitive governmental and infrastructure systems, potentially leading to significant confidentiality and operational impacts. No specific affected software versions or vulnerabilities have been disclosed, and no known exploits are currently reported in the wild. The threat is assessed as critical due to the broad scope, high-value targets, and potential for severe disruption. European organizations, especially those involved in government and critical infrastructure sectors, face heightened risk. Mitigation requires enhanced network monitoring, threat intelligence sharing, and rigorous access controls tailored to detect and prevent APT intrusions. Countries with significant government and critical infrastructure digital footprints, such as Germany, France, the UK, Italy, and the Netherlands, are most likely to be affected given their strategic importance and historical targeting patterns. The threat's complexity and stealthy nature necessitate proactive defense measures beyond standard patching and antivirus solutions.
A new hacking group called Punishing Owl has emerged, targeting Russian critical infrastructure. Their first attack on December 12, 2025, compromised a Russian state security agency, leaking internal documents. The group used DNS manipulation, created fake subdomains, and sent phishing emails to the victim's partners. They employed a PowerShell stealer called ZipWhisper to exfiltrate browser data. Punishing Owl's attacks are politically motivated and focus exclusively on Russian targets, including government agencies, scientific institutions, and IT organizations. The group has established a presence on cybercriminal forums and social media, likely operating from Kazakhstan. Experts predict this group will continue to be a persistent threat in the Russian cyberspace.
While both cyberwar and cyberwarfare will increase through 2026, cyberwarfare is likely to increase more dramatically. We hope it will never boil over – but we should be aware of the possibility and its consequences. The post Cyber Insights 2026: Cyberwar and Rising Nation State Threats appeared first on SecurityWeek .
A massive malware campaign dubbed ClawHavoc has been uncovered in the ClawHub marketplace, targeting OpenClaw bots and their users. An AI bot named Alex, working with security researcher Oren Yomtov, discovered 341 malicious skills, including 335 from a single campaign. The malware, identified as Atomic Stealer (AMOS), uses sophisticated techniques to evade detection and steal sensitive data. The attack exploits users' trust in AI assistants, potentially compromising personal and financial information. In response, a new tool called Clawdex has been developed to help bots and users scan for malicious skills before installation.
Last week, a new AI agent framework was introduced to automate "live". It targets office work in particular, focusing on messaging and interacting with systems. The tool has gone viral not so much because of its features, which are similar to those of other agent frameworks, but because of a stream of security oversights in its design.
Infostealer threats are expanding beyond Windows, targeting macOS and leveraging cross-platform languages like Python. Recent campaigns use social engineering to deploy macOS-specific infostealers such as DigitStealer, MacSync, and AMOS. These stealers use fileless execution and native macOS utilities to harvest credentials and sensitive data. Python-based stealers are also on the rise, allowing attackers to quickly adapt and target diverse environments. Additionally, threat actors are abusing trusted platforms like WhatsApp and PDF converter tools to distribute malware such as Eternidade Stealer. These evolving threats blend into legitimate ecosystems and evade conventional defenses, posing significant risks to organizations across various operating systems and delivery channels.
Yesterday, a single IP address (%%ip:204.76.203.210%%) scanned a number of our sensors for what looks like an anthropic API node. The IP address is known to be a Tor exit node.
The likely state-sponsored threat actor had access to the hosting provider for months and targeted only certain Notepad++ customers. The post Notepad++ Supply Chain Hack Conducted by China via Hosting Provider appeared first on SecurityWeek .
Among them, 23,000 hosts were persistently responsible for the majority of activity observed over 293 days of scanning. The post 175,000 Exposed Ollama Hosts Could Enable LLM Abuse appeared first on SecurityWeek .
An LLMjacking operation has been targeting exposed LLMs and MCPs at scale, for commercial monetization. The post LLMs Hijacked, Monetized in ‘Operation Bizarre Bazaar’ appeared first on SecurityWeek .
The fastest way to squander the promise of AI is to mistake automation for assurance, and novelty for resilience. The post Why We Can’t Let AI Take the Wheel of Cyber Defense appeared first on SecurityWeek .
Quantum computers are coming, with a potential computing power almost beyond comprehension. The post Cyber Insights 2026: Quantum Computing and the Potential Synergy With Advanced AI appeared first on SecurityWeek .
The HoneyMyte APT group has enhanced its toolset with an updated CoolClient backdoor and new data stealing capabilities. The group targeted government entities in Asia and Europe, particularly Southeast Asia. CoolClient now features clipboard monitoring, HTTP proxy credential sniffing, and plugin support for extended functionality. HoneyMyte also deployed browser login data stealers and document theft scripts. The campaign's focus has shifted towards active surveillance, including keylogging, clipboard data collection, and proxy credential harvesting. Organizations are advised to remain vigilant against HoneyMyte's evolving toolkit, which includes CoolClient, PlugX, ToneShell, Qreverse, and LuminousMoth malware families.
A new study shows that verse-based prompts can slash the effectiveness of AI safety constraints. We’re breaking down an experiment involving 25 language models and its key takeaways.
The Contagious Interview campaign, attributed to North Korea, continues to target software developers through fake recruitment schemes. A new technique in their arsenal leverages Microsoft Visual Studio Code task files to execute malicious code when a project is opened. The report documents observations of this vector, presents GitHub-based discovery methods, highlights findings including a new malicious NPM package, and outlines detection opportunities. The campaign exploits VS Code's Task feature, using the runOptions property to automatically execute malicious shell commands when a workspace is opened. Various obfuscation techniques are employed, including hiding commands with whitespace and masquerading payloads as image or font files.
A team of researchers from the Graz University of Technology in Austria has revived page Linux page cache attacks. The post Old Attack, New Speed: Researchers Optimize Page Cache Exploits appeared first on SecurityWeek .
The title of this diary is perhaps a bit catchy but the question is important. I don't consider myself as a good developer. That's not my day job and I'm writing code to improve my daily tasks. I like to say “I'm writing sh*ty code! It works for me, no warranty that it will for for you”. Today, most of my code (the skeleton of the program) is generated by AI, probably like most of you.
