A new malicious campaign dubbed DOUBLECUP exploits the ClickFix technique and cached PNGs to distribute two malware families: CountLoader and DeviceManager RAT. Anatomy of a two-stage attack.
3 h ago 24 3
A flaw in the random number generator of the COLDCARD hardware wallet could have allowed the derivation and draining of thousands of wallets—one of the largest documented thefts on "cold" hardware.
yesterday 9 3
Unit 42 (Palo Alto) documents a real-world case of an autonomous LLM agent: a single instruction sent via Telegram was enough to scan the internet, select a public exploit, and launch an attack—without any further human intervention.
Jul 31, 2026 18 3
Anthropic ran Claude Mythos Preview as a semi-autonomous agent on cryptographic primitives: effective keysize of HAWK-256 dropped from 2^64 to 2^38, 7-round AES attack accelerated ×200 to ×800, LEA 13-round key recovery in less than an hour. Two real CVE (OpenSSL, wolfSSL) along the way. Cost: ~$100,000 API per subject.
Jul 28, 2026 20 4
CVE-2026-54121, CVSS 8.8. A domain account with no privileges is sufficient to obtain an AD CS-signed certificate for the identity of a Domain Controller, then authenticate via PKINIT and extract krbtgt via DCSync. Patched on July 14, 2026, public exploit since today.
Jul 24, 2026 3 1
After the Hugging Face breach linked to an autonomous AI agent system, The Register reframes the debate: the problem isn't the agent itself, but the stacking of permissions, hardcoded credentials, and undefined tasks entrusted to it. New piece in the "ai-agents-threat" thread.
Jul 24, 2026 18 3
Researchers demonstrate that a local attacker can intercept and replace a macOS application during its download, before Gatekeeper has finished its job. Apple, informed, downplays it. Details of a "dumb and nasty" attack that highlights a gray area in the macOS trust model.
Jul 24, 2026 5 2
Socket.dev documented a campaign where 583 GitHub Actions workflows slipped into 10 PHP packages scan the Internet for cPanel/WHM instances vulnerable to the CVE-2026-41940 authentication bypass - to steal almost all available credentials. Approximately 6,100 workflows bear the campaign's signature on GitHub.
Jul 23, 2026 18 3
The Swiss train manufacturer Stadler Rail has been targeted by the Everest group via a data exchange platform with a supplier. It refuses to pay the 10 M CHF (~12.3 M$) demanded - and at this hour, Everest has not published anything on its leak site, which is unusual.
Jul 23, 2026 5 2
Zenity Labs publicly disclosed on July 23, 2026 a vulnerability in the ChatGPT agent builder that allowed the silent installation of a malicious agent inheriting all of the victim's connectors (Outlook, Slack, Drive...). Fixed on June 9 - but the vulnerability class remains to be monitored.
Jul 23, 2026 7 2
A flaw in the Azure DevOps MCP (Model Context Protocol) integration allows an invisible comment in a pull request to hijack an LLM agent tasked with reviewing it.
Jul 23, 2026 7 2
A flaw in the Adobe Acrobat browser extension allows malicious sites to read the victim's WhatsApp Web data - a classic Same-Origin Policy bypass via a privileged extension.
Jul 23, 2026 18 3
