Cloud Security
AWS attaches a quarantine policy to keys it detects as leaked, but Corey Quinn argues in The Register that the deny-list still lets an attacker assume other roles, destroy audit logs, read secrets, and lock storage. Treat a leaked key as a full compromise.
Artificial Intelligence (AI)
An independent Guidelight assessment graded five frontier AI labs on six control practices and found few publicly documented plans for containing a rogue model. The scores measure public disclosure, not private preparedness, but only OpenAI cleared even a partial mark on containment.
Industrial Control Systems (ICS)
New Nozomi Networks research, unveiled at Black Hat, shows how unprotected TSN (Time-Sensitive Networking) protocols could let attackers disrupt or manipulate physical processes in OT networks. It is lab research, not an active campaign, and segmentation is the first defense.
Policy & Government
The U.S. Justice Department says TikTok and ByteDance will pay $400 million to resolve a 2024 child-privacy lawsuit alleging COPPA violations. The deal pays $300 million now and $100 million once a court vacates an earlier Musical.ly consent decree.
Supply Chain Attack
Wiz researchers linked the crates.io compromise of three Rust crates to recent North Korean supply-chain operations, pointing to command-and-control infrastructure shared with the Mastra and axios npm attacks. It is a strong lead built on overlap, not a confirmed verdict.
Phishing
Researchers at Abnormal documented iAuthFlow V2, a $10,000 phishing toolkit that registers an attacker-controlled passkey on a compromised account. Because a passkey is a credential, not a password token, it survives the reset and session revocation that normally evict a phisher.
Supply Chain Attack
Trend Micro's TrendAI team disclosed 14 trojanized npm packages that pose as calendar and streak utilities while dropping RedC2 4.0, a cross-platform Linux backdoor with AI-assisted command-and-control. Here is what the disclosure confirms, what it does not, and what defenders should audit now.
Check Point Research disclosed that BTR.sys, Microsoft Defender's own legitimately signed remediation driver, can be turned into a kernel operation primitive for arbitrary file and registry changes across Windows 7 through Windows 11 25H2, with no external driver imported.
The Rust Project deleted malicious releases of arrayref, internment, and append-only-vec from crates.io after a compromised maintainer account added a typosquatted proc-macro1 dependency whose build script ran malware at compile time. Here is what security teams should check.
Cisco Talos published two reports on a Chinese-speaking cybercrime group, UAT-10147, that it says built agentic AI into its post-compromise operations, along with a companion cross-platform implant called SPECTRE. Here is what was disclosed, and what defenders should do about it.
Microsoft says a maximum-severity remote code execution flaw in Entra ID, CVE-2026-69836, was exploited in the wild before it fixed the service on its own infrastructure. No patch to install does not mean there is nothing for security teams to check.
The Hacker News has independently confirmed GitLab CVE-2026-19478, a CVSS 9.4 code injection flaw, is under active exploitation within days of disclosure, corroborating watchTowr. The unauthenticated GraphQL bug hits unpatched self-managed instances. Patch now.
A critical Elementor Pro flaw, CVE-2026-32475, lets unauthenticated visitors slip a PHP file past the Forms upload check and run code on the server. It is the second WordPress-plugin RCE disclosed in three days. Update to 4.2.2.
Our weekly wrap of stories that did not get a standalone piece: Iranian mega-theft charges, a pre-auth Citrix NetScaler bypass, Rapid7's record vulnerability quarter, fresh browser fixes, Grandoreiro in Mexico, an EDR-killing malware cocktail, and a no-filter AI service.
Threat actors began exploiting GitLab CVE-2026-19478 roughly two days after disclosure, according to WatchTowr. The critical, unauthenticated GraphQL flaw lets attackers modify or delete public projects and user data on unpatched self-managed instances. Fixed builds shipped August 17.
Researchers at the University of Massachusetts Amherst showed that a contactless credit card can keep making payments past its printed expiration date, even after a replacement arrives. They call it the Zombie Card attack, presented at USENIX Security 2026.
An engineer asked an AI coding agent for a package and it named one that was malware. Only a company rule to review any AI-recommended code on GitHub first caught the slop-squatting attempt before install.
Five US agencies, led by NSA, FBI and CISA, warn that an active campaign is using AI-assisted development to build tools targeting internet-exposed Siemens S7 Series PLCs, in what may be the first US-agency flag of AI-assisted tradecraft against OT.
OpenAI halted frontier reinforcement-learning training for two weeks and says expanded chain-of-thought monitoring will raise overhead about 20% on some workloads. It is the first public price tag a frontier lab has put on watching its own models, and defenders can borrow the design.