Bản tin hằng ngày về các lỗ hổng, vụ vi phạm và kỹ thuật tấn công đang định hình AI doanh nghiệp — và nơi quản trị AI chặn chúng trước khi đến production.
Các bài viết được đăng bằng tiếng Anh.
Claroc found 35,578 of 37,571 live Grok-built apps (94.7%) open without sign-in, some holding patient data and plain-text passwords. What happened, what to check.
Microsoft's November 2026 Patch Tuesday fixed a record 1,235 flaws, including CVE-2026-124194, an exploited Outlook spoofing zero-day that leaks NTLM hashes. What it does and what to check.
N-able took hosted N-central offline after a researcher published proof-of-concept exploits for 30+ alleged zero-days, and advised on-premises customers to shut servers down. What is known and what to check.
QNAP says NAS devices are being compromised through the Plex2Shell flaw chain, and Synology warns of C3RB3R ransomware hitting outdated DSM. Versions, advice, and how to find exposed NAS devices.
SonicWall fixed seven SonicOS flaws, including CVE-2026-19868, an SSL-VPN SAML authentication bypass under exploitation. Affected versions, fixes, and how to review VPN logins.
The *sshole flaw lets an oversized password log in to password-disabled accounts on musl-based OpenSSH servers such as Alpine. Conditions, the 10.6p2 fix, and what to check.
DigiCert disclosed that a February 2025 server migration exposed an internal system, and an unauthorized account read archived customer data including ID documents. What happened and how to triage the same signal.
mSeven Software says suspicious activity hit the mSecure password manager, and support replies reportedly describe a stolen encrypted cloud database. What is known, what is not, and a response checklist.
VulnCheck saw attackers exploit Node-RED CVE-2026-58115 (CVSS 10) from 30 October, planting a Linux bot and webshells. Siemens SIMATIC IoT2050 devices are affected. What to check.
CrowdStrike links Vapor Panda to AI-agent attacks on 200+ organizations in 14 countries, and DIVD logged an ARTEX agent reaching root in 12 minutes. What machine-speed intrusions look like and how to detect them.
CISA lists GitLab CVE-2026-97229 as exploited; Alpine Linux upgraded yet stayed exposed because Ruby needed patching too. About 2,800 CE instances are public.
WIRED reports an Accenture contractor connected a personal Claude Code agent to a Pentagon CUI system using their own credentials. It ran for about two months and sent passwords and a network map off-network. What that means for AI agent governance.
North Korea's TraderTraitor group is blamed for the record $6.3 billion Binance theft. The entry point was a compromised engineer laptop, and staging began 36 hours before the money moved. What the detection window looked like.
Dream reports a PHP2Shell attack chain that moved from exposed firewall management to an LDAP service account across 25 servers. How the pivot works and what to detect.
CISA added the Node.js Brotli flaw CVE-2026-111497 and the Adobe Commerce PolyShell flaw CVE-2026-48356 to its KEV catalog on 27 October 2026. Patches exist; here is why that is not enough.
Google's Threat Intelligence Group says criminals used an AI model to build a zero-day 2FA bypass for mass exploitation, and disrupted it before launch. What it means for patching.
Obiguard biến những sự cố bạn đọc thành chính sách được thực thi — rào chắn bảo vệ, ánh xạ tuân thủ và chuỗi kiểm toán cho mọi yêu cầu AI mà tổ chức của bạn thực hiện.
Trao đổi với chúng tôi →