
Sign up to save your podcasts
Or


Two critical RCE vulnerabilities in PHP affect web servers globally. Mirai botnet operators are already incorporating PHP2Shell exploits with ASLR bypass techniques documented. No successful compromises confirmed yet — but patch immediately. Kaspersky SIEM rules and Unit 42 threat brief available. SecurityWeek: Mirai botnet joins PHP2Shell attacks.
RCE vulnerability in Brotli decoders affecting Node.js, Python, and other runtimes. Any application decompressing Brotli data from untrusted sources may be vulnerable. Check dependency chains.
Kaspersky documents a new Lazarus Group campaign targeting South Korea following Operation SyncHole. Updated malware chain combining old and new implants. Nation-state threat actor with geopolitical targeting objectives.
Kaspersky GERT details a Middle East manufacturing incident where threat actors obtained domain admin access and deployed ransomware estate-wide via malicious Group Policy Objects. Audit GPO change permissions.
ICANN CTO confirms accelerating work to prepare DNS infrastructure for post-quantum cryptography. Internet governance bodies pushing to ensure DNS remains secure once quantum computers can challenge conventional encryption.
No new Known Exploited Vulnerabilities added to the CISA KEV catalogue in the past 48 hours.
41 victims · 11 groups
N0n (11) — continuing volume campaign with immediate data publication
Qilin (8)
Krybit (6)
Emperador (5)
The Gentlemen (3)
Falcon (2)
Whitearmor (2) — Partido Liberal 🇦🇷 | Hisense 🇨🇳
HYFLOCK (1) — ⚠️ Aesthetic Dental Designs 🇺🇸 — Healthcare
Anubis (1) — Newcomer Cremations 🇺🇸
Lamashtu (1) — Vins Viallet 🇫🇷 — Retail
Panzer (1) — Yodingenierie 🇧🇯 — Professional Services
BlackX UEFI Bootkit — Active espionage bootkit exploiting Baton Drop Secure Boot bypass. Survives OS reinstalls. Kills Kaspersky and Avast.
Autonomous AI Attack Agent ("Skynet Mode") — First wild-deployment of AI-on-AI offensive tooling. Runs 100,000-step attacks with zero human approval.
Ransomware: Qilin/Castle Howard, Whitearmor/JB Hunt, N0n healthcare, Emperador/AB InBev
Telegram Desktop CVE-2026-107181 — IPC injection, session exfiltration. Patch: v7.2.9.
EDR-Killing Driver Market — Commoditised kernel drivers, $1,500/driver.
HTTP2Shell (CVE-2026-111497) — Python HTTP/2 library attack escalation.
BlackX UEFI Bootkit — Active espionage bootkit exploiting the Baton Drop Secure Boot bypass. Survives OS reinstalls. Kills Kaspersky and Avast on load.
Autonomous AI Attack Agent ("Skynet Mode") — Symantec documented first wild-deployment of AI-on-AI offensive tooling: agent hijacks exposed AI endpoints to run up to 100,000-step attacks with zero human approval.
Ransomware: Qilin, Whitearmor, N0n, Emperador — Qilin posts Castle Howard Estate (UK) among 14+ victims. Whitearmor claims JB Hunt ($5M) and Whirlpool. Emperador posts AB InBev ($1M). N0n publishes healthcare data immediately across 6+ victims.
Telegram Desktop CVE-2026-107181 — High-severity IPC injection via malformed tg:// links, exfiltrates session data. Patch: version 7.2.9.
EDR-Killing Driver Market — Commoditised kernel drivers with memory R/W targeting EDR/AV processes, $1,500/driver with weekly new variants.
Samsung TV Root + HTTP2Shell (CVE-2026-111497) — Samsung TV AI silently exfiltrating to OpenAI; root shell demonstrated. Python HTTP/2 library attack escalation.
Firmware persistence moves beyond nation-state. AI becomes an attacker operational capability. Ransomware bifurcates into negotiators vs. immediate publishers.
NCSC blog by Toby W covers 7 safeguard areas for agentic AI deployment: autonomy scoping, model safeguard assessment, additional controls, threat modelling, prompt constraints, oversight calibration, and sandboxing. Treat agentic AI logging as a SOC function.
Four breaches added this week: Crunchbase (559,908 accounts, professional data), Sharecare (318,185, health and government IDs), Gîtes de France (117,574, booking details), Assurance Retraite (70,372, French pension and financial data).
From the publisher's feed