Attackers hijacked the npm account of a former contributor to Mastra, a popular open-source framework for building AI applications, and in an 88-minute automated burst republished 144 packages under the @mastra scope with a hidden malicious dependency. The poisoned dependency, a fake clone of a date library, runs at install time: it disables TLS checks, downloads a second-stage cryptocurrency-stealing trojan, runs it as a detached process, and deletes itself. Because @mastra/core alone sees over 900,000 weekly downloads and the payload fires on install, anyone who installed an affected version since June 16 could be compromised before importing anything. npm has pulled the malicious versions.
Eastman Kodak has confirmed that an unauthorized third party gained temporary access to a limited amount of company data, after the extortion group ShinyHunters listed the firm on its dark-web leak site. ShinyHunters claims it stole more than 2.2 million records containing customer personal information and internal corporate data, and set a leak deadline of June 18, though it has released no proof and Kodak has not verified the figure. Kodak, now mainly a B2B manufacturing and technology company, says it engaged outside experts and law enforcement and sees no threat to operations. The breach fits ShinyHunters' prolific 2026 data-theft campaign.
Symantec reports that DragonForce ransomware operators stayed hidden inside a major US services firm's network for up to two months by disguising their command-and-control traffic as ordinary Microsoft Teams activity. A new Go-based backdoor, Backdoor.Turn, grabs an anonymous Teams visitor token, routes through a legitimate Microsoft Teams relay server, and then tunnels to the attackers' real server, so defenders watching the network only see connections to genuine Microsoft infrastructure. It is the first known malware to abuse Teams relay servers this way. The attackers also used a custom malicious driver to disable defenses, and installed the backdoor after deploying ransomware, suggesting they kept access for a return visit or to resell.
Have I Been Pwned has added 248,235 accounts from the March breach of CFGI, a US accounting and financial-advisory firm that works closely with corporate finance teams at mid-market and Fortune 500 companies. The extortion group ShinyHunters claimed the intrusion, posting hundreds of thousands of records including names, emails, phone numbers, and home addresses, along with internal corporate documents and identity-system metadata. Because CFGI sits inside its clients' finance functions, the stolen contact and relationship data is unusually useful for convincing business email compromise and client-impersonation scams aimed at authorizing fraudulent payments.
CISA has added a LiteSpeed cPanel plugin flaw to its known-exploited list and given federal agencies until June 18 to patch. The bug (CVE-2026-54420, rated 8.5) lets a user who already has FTP or web-shell access on a shared hosting server escalate to root by abusing how the plugin follows symbolic links, on servers running CloudLinux or CageFS. On multi-tenant hosting that turns one compromised account into full control of the whole server and every site on it. Namecheap reported it after spotting suspicious activity, and LiteSpeed flagged active exploitation in early June. The fix is LiteSpeed WHM Plugin 5.3.2.1 with cPanel plugin 2.4.8.
Threat-intelligence firm Defused reports that attackers are now exploiting three critical flaws in Fortinet's FortiSandbox, the appliance other Fortinet products rely on to judge whether files are malicious. Two (CVE-2026-39813, a JRPC API path traversal that bypasses authentication, and CVE-2026-39808, an unauthenticated command-injection that runs code as root) were patched in April; the third (CVE-2026-25089) only last week. All are unauthenticated and rated critical. Compromising a sandbox is especially dangerous because attackers can make it wave real malware through as clean. Notably, the exploit for one flaw appears to have been generated with AI and is likely faulty, yet attackers are trying it anyway.
Palo Alto's Unit 42 disclosed a flaw, nicknamed Pickle in the Middle, in Google Cloud's Vertex AI SDK for Python that let an attacker with no access to a victim's project hijack their machine-learning model uploads and run code across tenant boundaries. When a model was uploaded without a custom staging bucket, the SDK generated a predictable storage bucket name from the project ID and region and failed to verify ownership, so an attacker could pre-create that bucket, receive the victim's model, and swap in a malicious one that executes on deployment. Google fully fixed it in SDK version 1.148.0 in April; Unit 42 saw no exploitation in the wild.
iRhythm, the US digital-health company behind the Zio wearable heart monitor, has told regulators that attackers stole patient data in a breach it considers material. In an SEC filing, the company said it detected unauthorized activity on June 8 in third-party-hosted business applications, accessed through a social-engineering attack, and received an extortion demand the next day from a threat actor claiming to hold proprietary data, protected health information, and other personal data. iRhythm says its clinical systems, medical devices, patient safety, and operations were not affected, with no payment-card or financial data involved. No ransomware group has publicly claimed the attack, and the number of affected people is not yet known.
ESET has found two previously unknown Windows versions of SprySOCKS, a backdoor until now seen only on Linux, attributed to the China-aligned espionage group FishMonger (also called Earth Lusca and linked to the i-Soon contractor). One variant loads two encrypted kernel drivers that hide the malware's processes, files, registry keys, and network connections, and divert command traffic through a random TCP port so the real listening port never shows. It keeps the Linux version's 30-plus commands and hardcoded command-and-control setup. ESET tied the activity to attacks in 2023 and 2024, mostly against government bodies in Honduras, Taiwan, Thailand, and Pakistan, with the group historically gaining entry through unpatched public-facing servers.
Zimperium's zLabs has documented Rokarolla, a new Android banking trojan that targets 217 banking and cryptocurrency apps and accepts 137 remote commands, giving an operator near-total control of an infected phone. It lifts lock-screen PINs, reads and sends text messages to grab one-time codes, rewrites the clipboard to redirect cryptocurrency payments, and disables Google Play Protect. It spreads through malicious websites posing as popular apps like TikTok and Chrome, starting with a dropper disguised as Google Play Protect that abuses Accessibility permissions. The actual theft uses fake login overlays placed on top of real banking apps, and surveillance relies on quiet Accessibility screenshots.
TrustStrike Labs