Braintrust, an AI evaluation and observability platform recently valued at $800 million, confirmed Tuesday that an unauthorized actor accessed one of its AWS accounts on May 4. The breached account held org-level API keys that customers store with Braintrust to access OpenAI, Anthropic, and other AI providers. Braintrust has confirmed exposure of one customer and is investigating three more reporting suspicious AI-provider usage spikes. The pattern - a relatively small AI infrastructure provider becoming a credential warehouse for downstream customers - is what Nudge Security's Jaime Blasco called 'the new shape of supply chain risk.'
cPanel released patches Friday for three new vulnerabilities. The two worst (CVE-2026-29202 and CVE-2026-29203, both CVSS 8.8) let authenticated users execute arbitrary Perl code through the create_user API or escalate privileges via unsafe symlink chmod. The third (CVE-2026-29201, CVSS 4.3) lets authenticated users read arbitrary files. No exploitation observed yet. The disclosure lands while attackers are still mass-exploiting CVE-2026-41940 to deploy 'Sorry' ransomware against cPanel hosts, including a wave targeting government agencies and MSPs (covered May 5). Hosting providers face a compounding patch burden.
Researcher Hyunwoo Kim disclosed Dirty Frag yesterday after an unrelated third party broke the embargo five days early. The flaw chains two Linux kernel page-cache write bugs (xfrm-ESP and RxRPC) to give any local user root access on every major distribution - Ubuntu, RHEL, CentOS Stream, AlmaLinux, openSUSE Tumbleweed, Fedora. Like Dirty Pipe and last week's Copy Fail, it's a deterministic logic bug with no race condition required and no kernel panic on failure. PoC is public on GitHub. The ESP variant patch was merged into the netdev tree on May 7 but distribution kernels remain unpatched. No CVE assigned yet because the embargo broke early.
Update on the Instructure breach we covered May 4: ShinyHunters has shifted from extorting Instructure itself to extorting individual schools and universities with their own Canvas data. BleepingComputer and Krebs on Security report that 8,800+ institutions have received direct ransom demands referencing real student records, teacher accounts, and gradebook data from their own Canvas tenants. The campaign mirrors the 2025 PowerSchool aftermath. Some schools are receiving demands sized to the institution. Krebs notes affected schools are scrambling to comply with state student-privacy laws while negotiating with attackers.
Update on the Trellix breach we covered May 2: RansomHouse claimed the attack on its leak site Thursday and published screenshots that suggest the intrusion reached well beyond the source code repository Trellix originally acknowledged. Cybernews researchers reviewed the dumped images and identified internal dashboards for VMware vCenter, Rubrik backup, and Dell EMC storage - the systems that hold backups, credentials, and virtual machine images for the entire company. RansomHouse says the intrusion happened April 17 and resulted in data encryption. Trellix told BleepingComputer it's 'aware of claims of responsibility' and looking into them. RansomHouse currently lists 170+ victims on its Tor leak site.
NVIDIA confirmed Friday that a third-party GeForce NOW Alliance partner based in Armenia (GFN.am) was breached. The hacker, using the ShinyHunters handle on BreachForums, claims to have stolen names, email addresses, dates of birth, membership status, and 2FA enrollment status of millions of users - and is selling the database for $100,000. NVIDIA says its own systems are unaffected and the regional partner is notifying impacted users. The actor is suspected to be a ShinyHunters impersonator rather than the original gang. The partner serves users in Armenia, Georgia, Kazakhstan, Moldova, Ukraine, and Uzbekistan.
Group-IB and Flare disclosed PamDOORa, a new Linux backdoor for sale on the Russian-speaking Rehub cybercrime forum at $900 (down from $1,600). PamDOORa hijacks the Linux Pluggable Authentication Module (PAM) framework that handles SSH logins - so it intercepts every legitimate user's password as they authenticate, before any application-level logging fires. The backdoor injects a malicious pam_linux.so module into the authentication stack rather than replacing files. It also tampers with lastlog, btmp, utmp, and wtmp to erase attacker login traces - meaning incident response teams who SSH in to investigate will have their own credentials silently stolen. Group-IB notes the abuse method is not yet in MITRE ATT&CK.
ESET disclosed CallPhantom, a campaign of 28 fraudulent Android apps on Google Play that promised to reveal call histories, SMS records, and WhatsApp call logs for any phone number. Combined downloads: 7.3 million. After payment (weekly, monthly, or annual subscriptions up to $80), users receive fabricated phone numbers and names hardcoded into the apps. Targeting was India-focused (apps came pre-set with +91 country code and UPI integration via Google Pay, PhonePe, and Paytm) plus broader Asia-Pacific. Some apps embedded direct credit card forms, violating Play policy and making refunds harder. Google removed the 28 apps after ESET's report.
Update on the Head Mare campaign we covered April 28: Kaspersky now reports that BO Team (also known as Black Owl) and Head Mare appear to be coordinating cyber operations against Russian organizations, sharing command-and-control infrastructure on the same compromised hosts. The likely division of labor: Head Mare phishes for initial access, then BO Team takes over for malware deployment. BO Team has shifted from destructive attacks to covert espionage, and in Q1 2026 hit 20 Russian organizations across manufacturing, telecoms, and oil and gas. The group uses BrockenDoor and Remcos backdoors. Earlier BO Team campaigns hit a Russian drone supplier and the federal digital signature authority.
Researchers disclosed TCLBANKER, an Android banking trojan that adds worm-style self-propagation: once installed, it abuses Accessibility Services to read the victim's WhatsApp and Outlook contact lists and then send malicious download links to every contact as if from the victim. The malware targets banking and crypto-wallet apps with overlay screens that capture credentials, plus SMS-interception modules that grab one-time passcodes. Self-spreading via the victim's own messaging history defeats traditional URL-reputation controls. The campaign concentrates in Brazil, Spain, and Italy banking apps initially. Operators are renting access on Telegram for $1,500-3,000/month.
TrustStrike Labs