San Mateo, CA, February 10, 2025 — Stories, events, and developments that impacted the cybersecurity landscape last week, including emerging threats, policy changes, and industry responses.
Sites posing as DeepSeek steal data and crypto wallets
China’s DeepSeek AI model sent waves through the industry, rocketing it to the top of the charts and the forefront of internet searches. Threat actors have noticed and built phishing sites that impersonate the platform. Israeli company Memyco discovered 16 phony sites to trick users into downloading malicious software that steals login credentials or crypto. Memyco researchers believe their findings indicate a coordinated, agile campaign involving different threat actors. “Memcyco observed clusters of fake domains registered in waves, often adjusting their content and branding dynamically and in real-time, based on how DeepSeek’s website was being perceived and positioned in the market,” says Israel Mazin, CEO and co-founder of Memcyco. “Some sites even changed their attack methods based on these trends to cater to what would be most effective.” Read more.
Weaponized AI models on Hugging Face
Two malicious machine learning models have been discovered on Hugging Face, a popular platform for sharing AI products. Uncovered by Reversing Labs, the models contain malicious code but were not flagged as unsafe by Hugging Face. The models were found to be exploiting a novel malware distribution technique by abusing Pickle, “a Python module that serializes and deserializes ML model data – an action known as Pickling by the Hugging Face community.” Pickle files are unsafe and should only be used by trusted sources, but their convenience often causes users to overlook security concerns. The malicious models were found to be using broken Pickle files, leading researchers to believe that they were proof-of-concept versions as opposed to a final malicious model. Upon being informed of their presence, Hugging Face removed the malicious models and updated their scanning tools to flag broken Pickle files. Read more.
U.S. government customer of Israeli spyware maker
Israeli spyware maker Paragon, making headlines recently due to its wares being used in a campaign on WhatsApp, has confirmed to TechCrunch that one of its customers is the U.S. government. “Paragon licenses its technology to a select group of global democracies — principally, the United States and its allies,” the company’s executive chairman, John Fleming, said. He also noted that Paragon “requires that all users agree to terms and conditions that explicitly prohibit the illicit targeting of journalists and other civil society figures. We have a zero-tolerance policy against such targeting and will terminate our relationship with any customer that violates our terms of service.” No elaboration was provided regarding who it considers to be U.S. “allies” or whether or not the company investigates allegations of abuse. It was reported in 2024 that Paragon’s U.S. subsidiary had signed a $2 million contract with the U.S. Immigration and Customs Enforcement, but Paragon did not provide any comment at that time. Read more.
New Ransomware tactic offers rewards to employees
Researchers at GroupSense have uncovered a new strategy that ransomware attackers are using: using their ransom notes to solicit employees willing to share company data in return for money. One such note included the following details after the usual demand details: “If you help us find this company’s dirty laundry, you will be rewarded. You can tell your friends about us. If you or your friend hates his boss, write to us, and we will make him cry, and the real hero will get a reward from us.” Some such notes are offering “millions of dollars.” The tactic is new. As such there is little information related to whether or not any disgruntled employees have taken the bait. However, security researchers warn that cybercriminals have “no accountability” and there is no guarantee that a willing participant would get paid or that their involvement would remain a secret. Read more.
Crypto-stealing kits in Google and Apple Apps
A new campaign called SparkCat has been spotted, with attackers hiding crypto-stealing software development kits (SDK) in apps available in the Google Play and Apple App stores. Discovered by Kaspersky, this marks the first time that a stealer has been found in Apple’s highly guarded app marketplace. Kaspersky reports that there are 18 infected Android apps and 10 iOS apps, one of which has been downloaded more than 50,000 times, many of which are still available for download. The malicious Android apps employ “a malicious Java component called ‘Spark,’ disguised as an analytics module. It uses an encrypted configuration file stored on GitLab, which provides commands and operational updates.” To infect those using Apple’s App Store, “the framework has different names like ‘Gzip,’ ‘googleappsdk,’ or ‘stat.’ Also, it utilizes a Rust-based networking module called “im_net_sys” to handle communication with the command and control (C2) servers.” Read more.
Netgear addresses severe RCE vulnerability in routers
A critical security advisory detailing a severe remote code execution (RCE) vulnerability has been issued by Netgear. The flaw tracked as CVE-2025-25246, could allow threat actors to execute arbitrary code remotely and affect Netgear’s popular XR1000, XR1000v2, and XR500 routers. The flaw has received a 9.8 CVSS score and does not require interaction or authentication by the victim, making it particularly dangerous. Netgear is urging all users of affected router models to use the latest firmware available. Netgear has not detailed how attackers could exploit this flaw. Read more.
MacOS users targeted by North Korean fake job interviews
North Korean threat actors associated with a malicious campaign called Contagious Interview use fake job interviews to deploy FERRET, the name given to a cocktail of Apple macOS malware strains. The attacks start with the attackers posing as recruiters on LinkedIn, who urge their targets to take on a video assignment. “Targets are typically asked to communicate with an interviewer through a link that throws an error message and a request to install or update some required piece of software such as VCam or CameraAccess for virtual meetings,” SentinelOne researchers Phil Stokes and Tom Hegel said in their report. The end goal, like most North Korean-sponsored attacks, is financially motivated, with the threat actors dropping a stealer intended to drain victims’ MetaMask wallets of cryptocurrency. Read more.
Grubhub data breach exposes driver and customer data
Food delivery company Grubhub has experienced a data breach that they traced to a third-party service provider. “Upon discovery, we promptly launched an investigation, identifying unauthorized access to an account associated with this provider,” Grubhub said in a statement. “We immediately terminated the account’s access and removed the service provider from our systems altogether.” The breach is said to have allowed threat actors to access the personal data of customers, merchants, and drivers, including names, email addresses, phone numbers, and partial payment card information. Grubhub has not yet disclosed how many individuals have been affected by the breach or when it occurred. Read more.
WhatsApp disrupts spyware campaign targeting journalists
Messaging app WhatsApp has reported that it disrupted a spyware campaign tied to Israeli company Paragon that targeted journalists and activists. WhatsApp has sent a cease-and-desist to Paragon, likely with confidence, given WhatsApp’s recent legal victory against NSO Group, another notorious Israeli spyware developer. The campaign targeted ninety people across two dozen countries, in which the purveyors used groups and malicious PDF files to infect victims. “We’ve reached out directly to people who we believe were affected,” said a WhatsApp spokesperson. “This is the latest example of why spyware companies must be held accountable for their unlawful actions. WhatsApp will continue to protect people’s ability to communicate privately.” Read more.
China-linked patient monitors exfiltrate sensitive data
The FDA has urged medical facilities to disconnect any Contec patient monitoring equipment from the internet immediately, as a trio of vulnerabilities found within Contec CMS8000 devices could allow an attacker to execute code remotely, crash the machine, or steal patient information. “Once the patient monitor is connected to the internet, it begins gathering patient data, including personally identifiable and protected health information, and exfiltrating the data outside the health care delivery environment,” the FDA said. CISA’s fact sheet about this issue warns that the backdoor accessible on Contec CMS8000 devices is not associated with software updates and seems to be there exclusively for data harvesting. “The [back door] provides neither an integrity-checking mechanism nor version tracking of updates,” CISA said. “When the function is executed, files on the device are forcibly overwritten, preventing the end customer—such as a hospital—from maintaining awareness of what software is running on the device.” The data is believed to be sent to a “third-party university” associated with China. Read more.
More cybersecurity news
- Last week’s news
- More cybersecurity news
- All articles sponsored by NetworkTigers
