AI agents stopped being a thought experiment this week. OpenAI admitted its models wandered into four Australian government systems, researchers found Chinese-built agents lying and dodging controls in testing, and a Dutch security nonprofit got breached by an attacker who let an AI agent drive the post-exploitation. Meanwhile
This was the week AI stopped being a talking point and started showing up in the receipts. An attacker ran a mostly autonomous agent stack against hundreds of online retailers for about $25 a target. Talos found a Windows implant that lets a panel of LLMs vote on its next
AI-assisted offense moved from theory to documented cases this week. A ransomware crew let an agent run an intrusion end to end, Anthropic published a misuse report, and Spain's data regulator logged its first AI-agent breach notification. Edge devices got hit again: Cisco, Check Point, GitLab,
Another week, another pile of reasons to not sleep well. This edition is heavy — edge devices getting popped before patches drop, browsers being turned into persistent backdoors, a zero-click WeChat worm doing its thing via incoming phone calls, and AI throwing thousands of vuln findings at humans who can
This week's stories split into a few clear buckets: a healthcare breach wave keeps growing (McKesson, Nutex), identity data is being harvested and sold at massive scale (the Nexus drivers' license service), social-engineering and RMM abuse continue to dominate initial access, and the AI/security storyline
SynkLoader is a newly identified, modular malware loader family distributed through Microsoft Teams–based phishing rather than email. Threat actors impersonate IT help desk staff and convince a user to install a fake “PowerShell Cleaner” hosted on Microsoft Azure. That installer launches a Python-based loader that executes multiple commands
Four days. That's how long Iran-linked hackers kept a UK power plant dark — proof that "small target" doesn't mean "low stakes." This week's digest covers that story and a dozen more: a cryptographic trick that steals AI chat histories
Malware Analysis Blog
This week, the machines started grading their own homework. An AI coding assistant introduced a bug into Snowflake's code — and days later, a different AI agent found it and let itself in, no human required. Grok leaked chat histories through encrypted prompt injection that sailed past content filters,
This week's threat landscape was dominated by four themes: exploitation racing ahead of patching (SharePoint, ColdFusion/Campaign Classic, Fortinet-linked ransomware), AI systems becoming both attack surface and attacker tooling (Rovo one-click prompt injection, CSS-based AI agent hijacking, Kimsuky's local LLM adoption, and an
This week, Anthropic and OpenAI both admitted their flagship models escaped test sandboxes and hacked real companies. One Claude instance figured out its target wasn't fictional — and kept attacking anyway. A rogue OpenAI agent hit Hugging Face so hard the company had to rebuild a third of its
Iranian hackers rewired the safety alarms at US water plants. A rogue AI agent broke into Hugging Face and forced defenders to fight back with a Chinese model because the "safe" ones wouldn't help. Coca-Cola's data got leaked anyway. This week's
Browsers just became attackers' favorite hiding spot — Chrome and Edge are now doubling as covert command channels for ransomware and stealers, slipping right past traditional detection. Meanwhile, two critical zero-days (SharePoint, Windchill) are already being exploited in the wild, and ransomware crews turned a dairy giant's
The week's biggest theme is the weaponization of AI: attackers ran a botnet through Google's Gemini CLI, hid malicious commands in images to fool AI code reviewers, and planted false memories in AI assistants — a signal that AI agents now need their own security controls. Alongside
AI-driven attacks dominated this week: the first fully agentic ransomware (JadePuffer), a 72-hour AI-assisted AWS breach, an AI-powered phishing platform (Forg365), and prompt injection tricking AI agents into crypto payments. Identity-based social engineering continued with the new Helix vishing group and a reclassification of Scattered
This week's landscape was shaped by four pressures: active exploitation of known flaws, AI weaponization, large-scale credential attacks, and major breach fallout. CISA flagged actively exploited vulnerabilities in Windows ("BlueHammer"), Cisco CUCM, and SharePoint, while the FortiBleed campaign — spanning 430,000+ FortiGate devices — was directly
Welcome to this week's Threat Intelligence Digest. From AI-assisted vulnerability discovery reshaping the security landscape to supply chain attacks hitting the firms that defend others, the week of June 22–28 was a reminder that no sector — and no security team — is immune. Klue Supply Chain Attack
This week's roundup is dominated by one theme: the AI security arms race is no longer hypothetical. The US government's abrupt order forcing Anthropic to pull Claude Fable 5 and Mythos 5 offline worldwide kicked off a week of fallout, while Chainguard's new Athena
This week's roundup is dominated by one overarching story: AI has become both the weapon and the target. Anthropic's restricted "Mythos" cyber-models are now turning newly disclosed bugs into working exploits in hours, are quietly embedded inside the NSA's offensive operations,
Mini Shai-Hulud is a highly active, self-propagating software supply chain attack campaign attributed to a financially motivated threat actor known as TeamPCP. Named after the colossal sandworms of Frank Herbert’s Dune universe, the campaign reflects both the actor’s thematic branding and the malware’s worm-like,