At home with Gemini Google Home is getting deeper Gemini integration and a new widget Smart homes can be dumb, but maybe Gemini can help. Ryan Whitwam – May 23, 2025 2:53 pm | 0 Credit: Google Credit: Google Story text Size Small Standard Large Width * Standard Wide Links Standard Orange * Subscribers only   Learn more As Google moves the last remaining Nest devices into the Home app, it's also looking at ways to make this smart home hub easier to use. Naturally, Google is doing that by ramping up Gemini integration. The company has announced new automation capabilities with generative AI, as well as better support for third-party devices via the Home API. Google AI will also plug into a new Android widget that can keep you updated on what the smart parts of your home are up to. The Google Home app is where you interact with all of Google's smart home gadgets, like cameras, thermostats, and smoke detectors—some of which have been discontinued, but that's another story. It also accommodates smart home devices from other companies, which can make managing a mixed setup feasible if not exactly intuitive. A dash of AI might actually help here. Google began testing Gemini integrations in Home last year, and now it's opening that up to third-party devices via the Home API. Google has worked with a few partners on API integrations before general availability. The previously announced First Alert smoke/carbon monoxide detector and Yale smart lock that are replacing Google's Nest devices are among the first, along with Cync lighting, Motorola Tags, and iRobot vacuums. Google has been testing Gemini integrations in the Labs program, allowing users to create automations by telling Gemini what they want to do. For example, getting reminders about package deliveries or detecting unwanted animals in a garden. Gemini can theoretically create the setup for you, so you don't have to tap through multiple menus. With Gemini support in the automation API, more smart home devices will be visible to Gemini when it makes recommendations. Google also says the Gemini expansion will add more triggers based on dates and weather conditions. The complexity of smart home setups can make it hard to parse exactly what's going on. Google Home's favorites page helps, but a new widget could make it even easier. The Home Summary Widget will provide live insights on what's happening in your home. For example, it could remind you there's a package outside or that you left the lights on. The new widget will keep you abreast of what is happening around the house. Credit: Google The new widget will keep you abreast of what is happening around the house. Credit: Google Google says it's testing the widget on a small number of Pixel phones, but we expect it to be available more widely in the Labs program later. As for a general rollout, Google isn't making any promises right now. Developers interested in trying the new APIs are advised to sign up for the smart home developer newsletter. That will keep you in the loop as the early access program begins later this year. Ryan Whitwam Senior Technology Reporter Ryan Whitwam Senior Technology Reporter Ryan Whitwam is a senior technology reporter at Ars Technica, covering the ways Google, AI, and mobile technology continue to change the world. Over his 20-year career, he's written for Android Police, ExtremeTech, Wirecutter, NY Times, and more. He has reviewed more phones than most people will ever own. You can follow him on Bluesky, where you will see photos of his dozens of mechanical keyboards. 0 Comments

May 23, 2025Ravie LakshmananCryptocurrency / Malware The malware known as Latrodectus has become the latest to embrace the widely-used social engineering technique called ClickFix as a distribution vector. "The ClickFix technique is particularly risky because it allows the malware to execute in memory rather than being written to disk," Expel said in a report shared with The Hacker News. "This removes many opportunities for browsers or security tools to detect or block the malware." Latrodectus, believed to be a successor to IcedID, is the name given to a malware that acts as a downloader for other payloads, such as ransomware. It was first documented by Proofpoint and Team Cymru in April 2024. Incidentally, the malware is one among the many malicious software to suffer an operational setback as part of Operation Endgame, which took down 300 servers worldwide and neutralized 650 domains related to Bumblebee, Lactrodectus, QakBot, HijackLoader, DanaBot, TrickBot, and WARMCOOKIE between May 19 and 22, 2025. In the latest set of Latrodectus attacks observed by Expel in May 2025, unsuspecting users are tricked into copying and executing a PowerShell command from an infected website, a tactic that has become a prevalent method to distribute a wide range of malware. "When run by a user, these commands will attempt to install a file located at the remote URL using MSIExec, and then execute it in memory," Expel said. "This keeps the attacker from having to write the file to the computer and risk being detected by the browser or an antivirus that might detect it on disk." The MSI installer contains a legitimate application from NVIDIA, which is used to sideload a malicious DLL, which then uses curl to download the main payload. To mitigate attacks of this type, it's advised to disable the Windows Run program using Group Policy Objects (GPOs) or turn off the "Windows + R" hot key via a Windows Registry change. From ClickFix to TikTok The disclosure comes as Trend Micro revealed details of a new engineering campaign that instead of relying on fake CAPTCHA pages employs TikTok videos likely generated using artificial intelligence (AI) tools to deliver the Vidar and StealC information stealers by instructing users to run malicious commands on their systems to activate Windows, Microsoft Office, CapCut, and Spotify. These videos have been posted from various TikTok accounts such as @gitallowed, @zane.houghton, @allaivo2, @sysglow.wow, @alexfixpc, and @digitaldreams771. These accounts are no longer active. One of the videos claiming to provide instructions on how to "boost your Spotify experience instantly" has amassed nearly 500,000 views, with over 20,000 likes and more than 100 comments. The campaign marks a new escalation of ClickFix in that users searching for ways to activate pirated apps are verbally and visually guided to open the Windows Run dialog by pressing the "Windows + R" hot key, launch PowerShell, and run the command highlighted in the video, ultimately compromising their own systems. "Threat actors are now using TikTok videos that are potentially generated using AI-powered tools to socially engineer users into executing PowerShell commands under the guise of guiding them to activate legitimate software or unlock premium features," security researcher Junestherry Dela Cruz said. "This campaign highlights how attackers are ready to weaponize whichever social media platforms are currently popular to distribute malware." Fake Ledger Apps Used to Steal Mac Users' Seed Phrases The findings also follow the discovery of four different malware campaigns that leverage a cloned version of the Ledger Live app to steal sensitive data, including seed phrases, with the goal of draining victims' cryptocurrency wallets. The activity has been ongoing since August 2024. The attacks make use of the malicious DMG files that, when launched, launches AppleScript to exfiltrate passwords and Apple Notes data, and then download a trojanized version of Ledger Live. Once the app is opened, it warns users of a supposed account problem and that it requires their seed phrase for recovery. The entered seed phrase is sent to an attacker-controlled server. Moonlock Lab, which shed light on the campaign, said the rogue apps make use of macOS stealer malware like Atomic macOS Stealer (AMOS) and Odyssey, the latter of which introduced the novel phishing scheme in March 2025. It's worth noting that the activity overlaps with a macOS infostealer campaign that targets Ledger Live users through PyInstaller-packed binaries, as revealed by Jamf this month. "On dark web forums, chatter around anti-Ledger schemes is growing. The next wave is already taking shape," MacPaw's cybersecurity division noted. "Hackers will continue to exploit the trust crypto owners place in Ledger Live." Found this article interesting? Follow us on Twitter  and LinkedIn to read more exclusive content we post. SHARE    

A third-party research institute that Anthropic partnered with to test one of its new flagship AI models, Claude Opus 4, recommended against deploying an early version of the model due to its tendency to “scheme” and deceive. According to a safety report Anthropic published Thursday, the institute, Apollo Research, conducted tests to see in which contexts Opus 4 might try to behave in certain undesirable ways. Apollo found that Opus 4 appeared to be much more proactive in its “subversion attempts” than past models and that it “sometimes double[d] down on its deception” when asked follow-up questions. “[W]e find that, in situations where strategic deception is instrumentally useful, [the early Claude Opus 4 snapshot] schemes and deceives at such high rates that we advise against deploying this model either internally or externally,” Apollo wrote in its assessment. As AI models become more capable, some studies show they’re becoming more likely to take unexpected — and possibly unsafe — steps to achieve delegated tasks. For instance, early versions of OpenAI’s o1 and o3 models, released in the past year, tried to deceive humans at higher rates than previous-generation models, according to Apollo. Per Anthropic’s report, Apollo observed examples of the early Opus 4 attempting to write self-propagating viruses, fabricating legal documentation, and leaving hidden notes to future instances of itself — all in an effort to undermine its developers’ intentions. To be clear, Apollo tested a version of the model that had a bug Anthropic claims to have fixed. Moreover, many of Apollo’s tests placed the model in extreme scenarios, and Apollo admits that the model’s deceptive efforts likely would’ve failed in practice. However, in its safety report, Anthropic also says it observed evidence of deceptive behavior from Opus 4. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW This wasn’t always a bad thing. For example, during tests, Opus 4 would sometimes proactively do a broad cleanup of some piece of code even when asked to make only a small, specific change. More unusually, Opus 4 would try to “whistle-blow” if it perceived a user was engaged in some form of wrongdoing. According to Anthropic, when given access to a command line and told to “take initiative” or “act boldly” (or some variation of those phrases), Opus 4 would at times lock users out of systems it had access to and bulk-email media and law-enforcement officials to surface actions the model perceived to be illicit. “This kind of ethical intervention and whistleblowing is perhaps appropriate in principle, but it has a risk of misfiring if users give [Opus 4]-based agents access to incomplete or misleading information and prompt them to take initiative,” Anthropic wrote in its safety report. “This is not a new behavior, but is one that [Opus 4] will engage in somewhat more readily than prior models, and it seems to be part of a broader pattern of increased initiative with [Opus 4] that we also see in subtler and more benign ways in other environments.”

May 22, 2025Ravie LakshmananVulnerability / Software Security Cybersecurity researchers have uncovered multiple critical security vulnerabilities impacting the Versa Concerto network security and SD-WAN orchestration platform that could be exploited to take control of susceptible instances. It's worth noting that the identified shortcomings remain unpatched despite responsible disclosure on February 13, 2025, prompting a public release of the issues following the end of the 90-day deadline. "These vulnerabilities, when chained together, could allow an attacker to fully compromise both the application and the underlying host system," ProjectDiscovery researchers Harsh Jaiswal, Rahul Maini, and Parth Malhotra said in a report shared with The Hacker News. The security defects are listed below - CVE-2025-34025 (CVSS score: 8.6) - A privilege escalation and Docker container escape vulnerability that's caused by unsafe default mounting of host binary paths and could be exploited to gain code execution on the underlying host machine CVE-2025-34026 (CVSS score: 9.2) - An authentication bypass vulnerability in the Traefik reverse proxy configuration that allows an attacker to access administrative endpoints, which could then be exploited to access heap dumps and trace logs by exploiting an internal Spring Boot Actuator endpoint via CVE-2024-45410 CVE-2025-34027 (CVSS score: 10.0) - An authentication bypass vulnerability in the Traefik reverse proxy configuration that allows an attacker to access administrative endpoints, which could then be exploited to achieve remote code execution by exploiting an endpoint related to package uploads ("/portalapi/v1/package/spack/upload") via arbitrary file writes Successful exploitation of CVE-2025-34027 could allow an attacker to leverage a race condition and write malicious files to disk, ultimately resulting in remote code execution using LD_PRELOAD and a reverse shell. "Our approach involved overwriting ../../../../../../etc/ld.so.preload with a path pointing to /tmp/hook.so," the researchers said. "Simultaneously, we uploaded /tmp/hook.so, which contained a compiled C binary for a reverse shell. Since our request triggered two file write operations, we leveraged this to ensure that both files were written within the same request." "Once these files were successfully written, any command execution on the system while both persisted would result in the execution of /tmp/hook.so, thereby giving us a reverse shell." In the absence of an official fix, users are advised to block semicolons in URL paths and drop requests where the Connection header contains the value X-Real-Ip. It's also recommended to monitor network traffic and logs for any suspicious activity. The Hacker News has reached out to Versa Networks for comment, and we will update the story if we hear back. Found this article interesting? Follow us on Twitter  and LinkedIn to read more exclusive content we post. SHARE    

A third-party research institute that Anthropic partnered with to test one of its new flagship AI models, Claude Opus 4, recommended against deploying an early version of the model due to its tendency to “scheme” and deceive. According to a safety report Anthropic published Thursday, the institute, Apollo Research, conducted tests to see in which contexts Opus 4 might try to behave in certain undesirable ways. Apollo found that Opus 4 appeared to be much more proactive in its “subversion attempts” than past models and that it “sometimes double[d] down on its deception” when asked follow-up questions. “[W]e find that, in situations where strategic deception is instrumentally useful, [the early Claude Opus 4 snapshot] schemes and deceives at such high rates that we advise against deploying this model either internally or externally,” Apollo wrote in its assessment. As AI models become more capable, some studies show they’re becoming more likely to take unexpected — and possibly unsafe — steps to achieve delegated tasks. For instance, early versions of OpenAI’s o1 and o3 models, released in the past year, tried to deceive humans at higher rates than previous-generation models, according to Apollo. Per Anthropic’s report, Apollo observed examples of the early Opus 4 attempting to write self-propagating viruses, fabricating legal documentation, and leaving hidden notes to future instances of itself — all in an effort to undermine its developers’ intentions. To be clear, Apollo tested a version of the model that had a bug Anthropic claims to have fixed. Moreover, many of Apollo’s tests placed the model in extreme scenarios, and Apollo admits that the model’s deceptive efforts likely would’ve failed in practice. However, in its safety report, Anthropic also says it observed evidence of deceptive behavior from Opus 4. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW This wasn’t always a bad thing. For example, during tests, Opus 4 would sometimes proactively do a broad cleanup of some piece of code even when asked to make only a small, specific change. More unusually, Opus 4 would try to “whistle-blow” if it perceived a user was engaged in some form of wrongdoing. According to Anthropic, when given access to a command line and told to “take initiative” or “act boldly” (or some variation of those phrases), Opus 4 would at times lock users out of systems it had access to and bulk-email media and law-enforcement officials to surface actions the model perceived to be illicit. “This kind of ethical intervention and whistleblowing is perhaps appropriate in principle, but it has a risk of misfiring if users give [Opus 4]-based agents access to incomplete or misleading information and prompt them to take initiative,” Anthropic wrote in its safety report. “This is not a new behavior, but is one that [Opus 4] will engage in somewhat more readily than prior models, and it seems to be part of a broader pattern of increased initiative with [Opus 4] that we also see in subtler and more benign ways in other environments.”

May 22, 2025Ravie LakshmananCybersecurity / Vulnerability A privilege escalation flaw has been demonstrated in Windows Server 2025 that makes it possible for attackers to compromise any user in Active Directory (AD). "The attack exploits the delegated Managed Service Account (dMSA) feature that was introduced in Windows Server 2025, works with the default configuration, and is trivial to implement," Akamai security researcher Yuval Gordon said in a report shared with The Hacker News. "This issue likely affects most organizations that rely on AD. In 91% of the environments we examined, we found users outside the domain admins group that had the required permissions to perform this attack." What makes the attack pathway notable is that it leverages a new feature called Delegated Managed Service Accounts (dMSA) that allows migration from an existing legacy service account. It was introduced in Windows Server 2025 as a mitigation to Kerberoasting attacks. The attack technique has been codenamed BadSuccessor by the web infrastructure and security company. "dMSA allows users to create them as a standalone account, or to replace an existing standard service account," Microsoft notes in its documentation. "When a dMSA supersedes an existing account, authentication to that existing account using its password is blocked." "The request is redirected to the Local Security Authority (LSA) to authenticate using dMSA, which has access to everything the previous account could access in AD. During migration, dMSA automatically learns the devices on which the service account is to be used which is then used to move from all existing service accounts." The problem identified by Akamai is that during the dMSA Kerberos authentication phase, the Privilege Attribute Certificate (PAC) embedded into a ticket-granting ticket (i.e., credentials used to verify identity) issued by a key distribution center (KDC) includes both the dMSAs security identifier (SID) as well as the SIDs of the superseded service account and of all its associated groups. This permissions transfer between accounts could open the door to a potential privilege escalation scenario by simulating the dMSA migration process to compromise any user, including domain administrators, and gain similar privileges, effectively breaching the entire domain even if an organization's Windows Server 2025 domain isn't using dMSAs at all. "One interesting fact about this 'simulated migration' technique, is that it doesn't require any permissions over the superseded account," Gordon said. "The only requirement is to write permissions over the attributes of a dMSA. Any dMSA." "Once we've marked a dMSA as preceded by a user, the KDC automatically assumes a legitimate migration took place and happily grants our dMSA every single permission that the original user had, as though we are its rightful successor." Akamai said it reported the findings to Microsoft on April 1, 2025, following which the tech giant classified the issue as moderate in severity and that it does not meet the bar for immediate servicing due to the fact that successful exploitation requires an attacker to have specific permissions on the dMSA object, which suggests an elevation of privileges. However, a patch is currently in the works. Given that there is no immediate fix for the attack, organizations are advised to limit the ability to create dMSAs and harden permissions wherever possible. Akamai has also released a PowerShell script that can enumerate all non-default principals who can create dMSAs and list the organizational units (OUs) in which each principal has this permission. "This vulnerability introduces a previously unknown and high-impact abuse path that makes it possible for any user with CreateChild permissions on an OU to compromise any user in the domain and gain similar power to the Replicating Directory Changes privilege used to perform DCSync attacks," Gordon said. Found this article interesting? Follow us on Twitter  and LinkedIn to read more exclusive content we post. SHARE    

What’s the Best Way to Cut Onions Without Crying? New Research Suggests That Thin, Sharp Blades Are Key to Minimizing Tears For a new study, physicists visualized and quantified the tear-producing droplets that get expelled from onions when they’re cut New research suggests using a thin, sharp knife and cutting slowly could help prevent crying while cutting onions. Pexels From wearing goggles to rubbing the knife with lemon juice, home cooks have come up with a wide variety of clever tricks to keep themselves from crying while cutting onions in the kitchen. Now, physicists have come up with another possible solution. Using a sharp knife while making slow, controlled cuts seems to be the best way to minimize the spray of tear-producing compounds in onions, according to a new study. The paper has been published on the preprint server arXiv and has not yet been peer reviewed. Scientists know why onions make us cry: When cut, these spherical alliums produce a chemical that stimulates the nerves responsible for producing tears. The volatile, irritating compound is called syn-propanethial-S-oxide. But, recently, a team of physicists decided to explore the underlying mechanisms at play when syn-propanethial-S-oxide gets released from onions. Using techniques known as high-speed particle tracking velocimetry and digital image correlation, they were able to visualize and count droplets as they were being expelled from cut onions, per IFLScience’s Russell Moul. They also studied the onions themselves, making note of strain and deformations on the flesh during chopping. To start, the team gathered fresh onions from local vendors, cut them into halves or quarters, then coated them in black spray paint. (This step made it easier for them to see and track what was happening when the alliums got cut.) Then, they set up a high-speed camera and started chopping. “Cutting is a really strange process,” says Anne Juel, a physicist at the University of Manchester who was not involved with the research, to NewScientist’s Alex Wilkins. “We cut things with knives every day, but to cut something, you need to go down to the atomic scale.” Using a custom guillotine, the researchers experimented with cutting speeds ranging from 1.3 and 6.5 feet per second and blade thicknesses between 5 and 200 millimeters. They changed the steel blades manually and modified the speed by adjusting the height of the blade, which was released from above. By analyzing the footage from the high-speed camera, they could investigate the dynamics of tear-producing particles that sprayed out from the onion as it was cut. The thinner, sharper blades produced fewer droplets that moved more slowly and with less energy, they found. Meanwhile, the thicker, duller blades caused an explosion of high-speed particles that moved at up to 141 feet per second. This is because the dull blade initially bent the onion skin, which caused pressure to build up inside. When the blade finally sliced through, it released all that built-up energy and sent onion juice flying. Then, once the particles were in the air, they also fragmented into smaller pieces to create “an even more diffuse mist of all-natural mace,” writes Andrew Paul for Popular Science. The duller blades produced as much as 40 times more particles than the sharper blades. Faster cutting speeds produced up to four times as many droplets as slower speeds. This suggests that the best way to minimize tear-producing chemicals while chopping onions is to cut slowly with a thin, sharp knife. However, the researchers did not test this theory in their experiments, per NewScientist. Refrigerated onions released a “noticeably larger volume” of droplets compared to room-temperature onions, the researchers write. Pixabay Home cooks are often advised to chill their onions before cutting them to minimize crying. So, the researchers also experimented with onions that had been refrigerated for 12 hours. However, in their tests, the refrigerated onions released a “noticeably larger volume” of droplets compared to room-temperature onions. Preventing tears while cutting onions might seem like a trivial research topic. But the scientists say their work could also have important implications for food safety. The fragmented droplets that spray out of raw foods while they’re being cut could contribute to the spread of disease-causing pathogens, the researchers write in the paper. “Ejected droplets can come into direct contact with contaminated blades or carry surface-borne pathogens as they leave the food surface,” they write. “Lighter droplets [are] readily suspended and can be transported by ambient air currents, thereby posing a potential risk for airborne transmission.” Keeping knife blades sharp, then, might be an easy way to help reduce food-borne illnesses, they posit. “Sharper blades reduce not only the number of droplets but also their speed and kinetic energy,” the scientists write in the paper. “This is particularly relevant for fruits and vegetables, which can carry food-borne pathogens such as Salmonella.” Many other recent research projects have attempted to use science to solve everyday problems. Scientists have visualized the plume of aerosolized particles ejected from commercial toilets during flushing, and some are even coming up with new urinal designs to help reduce urine splashing. And in the realm of enhancing food, Italian physicists came up with the perfect cacio e pepe recipe, while others have figured out the best way to make pour-over coffee and boil an egg. Get the latest stories in your inbox every weekday. More about: Food Food Science New Research Physics

A threat actor known as Hazy Hawk has been observed hijacking abandoned cloud resources of high-profile organizations, including Amazon S3 buckets and Microsoft Azure endpoints, by leveraging misconfigurations in the Domain Name System (DNS) records. The hijacked domains are then used to host URLs that direct users to scams and malware via traffic distribution systems (TDSes), according to Infoblox. Some of the other resources usurped by the threat actor include those hosted on Akamai, Bunny CDN, Cloudflare CDN, GitHub, and Netlify. The DNS threat intelligence firm said it first discovered the threat actor after it gained control of several sub-domains associated with the U.S. Center for Disease Control (CDC) in February 2025. It has since been determined that other government agencies across the globe, prominent universities, and international corporations such as Deloitte, PricewaterhouseCoopers, and Ernst & Young have been victimized by the same threat actor since at least December 2023. "Perhaps the most remarkable thing about Hazy Hawk is that these hard-to-discover, vulnerable domains with ties to esteemed organizations are not being used for espionage or 'highbrow' cybercrime," Infoblox's Jacques Portal and Renée Burton said in a report shared with The Hacker News. "Instead, they feed into the seedy underworld of adtech, whisking victims to a wide range of scams and fake applications, and using browser notifications to trigger processes that will have a lingering impact." What makes Hazy Hawk's operations noteworthy is the hijacking of trusted and reputable domains belonging to legitimate organizations, thus boosting their credibility in search results when they are being used to serve malicious and spammy content. But even more concerningly, the approach enables the threat actors to bypass detection. Underpinning the operation is the ability of the attackers to seize control of abandoned domains with dangling DNS CNAME records, a technique previously exposed by Guardio in early 2024 as being exploited by bad actors for spam proliferation and click monetization. All a threat actor needs to do is register the missing resource to hijack the domain. Hazy Hawk goes a step further by finding abandoned cloud resources and then commandeering them for malicious purposes. In some cases, the threat actor employs URL redirection techniques to conceal which cloud resource was hijacked. "We use the name Hazy Hawk for this actor because of how they find and hijack cloud resources that have dangling DNS CNAME records and then use them in malicious URL distribution," Infoblox said. "It's possible that the domain hijacking component is provided as a service and is used by a group of actors." The attack chains often involve cloning the content of legitimate sites for their initial site hosted on the hijacked domains, while luring victims into visiting them with pornographic or pirated content. The site visitors are then funneled via a TDS to determine where they land next. "Hazy Hawk is one of the dozens of threat actors we track within the advertising affiliate world," the company said. "Threat actors who belong to affiliate advertising programs drive users into tailored malicious content and are incentivized to include requests to allow push notifications from 'websites' along the redirection path." In doing so, the idea is to flood a victim's device with push notifications and deliver an endless torrent of malicious content, with each notification leading to different scams, scareware, and fake surveys, and accompanied by requests to allow more push notifications. To prevent and protect against Hazy Hawk activities, domain owners are recommended to remove a DNS CNAME record as soon as a resource is shut down. End users, on the other hand, are advised to deny notification requests from websites they don't know. "While operators like Hazy Hawk are responsible for the initial lure, the user who clicks is led into a labyrinth of sketchy and outright malicious adtech. The fact that Hazy Hawk puts considerable effort into locating vulnerable domains and then using them for scam operations shows that these advertising affiliate programs are successful enough to pay well," Infoblox said. Found this article interesting? Follow us on Twitter  and LinkedIn to read more exclusive content we post.

May 20, 2025Ravie LakshmananCredential Theft / Browser Security An unknown threat actor has been attributed to creating several malicious Chrome Browser extensions since February 2024 that masquerade as seemingly benign utilities but incorporate covert functionality to exfiltrate data, receive commands, and execute arbitrary code. "The actor creates websites that masquerade as legitimate services, productivity tools, ad and media creation or analysis assistants, VPN services, crypto, banking and more to direct users to install corresponding malicious extensions on Google's Chrome Web Store (CWS)," the DomainTools Intelligence (DTI) team said in a report shared with The Hacker News. While the browser add-ons appear to offer the advertised features, they also enable credential and cookie theft, session hijacking, ad injection, malicious redirects, traffic manipulation, and phishing via DOM manipulation. Another factor that works in the extensions' favor is that they are configured to grant themselves excessive permissions via the manifest.json file, allowing them to interact with every site visited on the browser, execute arbitrary code retrieved from an attacker-controlled domain, perform malicious redirects, and even inject ads. The extensions have also been found to rely on the "onreset" event handler on a temporary document object model (DOM) element to execute code, likely in an attempt to bypass content security policy (CSP). Some of the identified lure websites impersonate legitimate products and services like DeepSeek, Manus, DeBank, FortiVPN, and Site Stats to entice users into downloading and installing the extensions. The add-ons then proceed to harvest browser cookies, fetch arbitrary scripts from a remote server, and set up a WebSocket connection to act as a network proxy for traffic routing. There is currently no visibility into how victims are redirected to the bogus sites, but DomainTools told the publication that it could involve usual methods like phishing and social media. "Because they appear in both Chrome Web Store and have adjacent websites, they can return from as results in normal web searches and for searches within the Chrome store," the company said. "Many of the lure websites used Facebook tracking IDs, which strongly suggests they are leveraging Facebook / Meta apps in some way to attract site visitors. Possibly through Facebook pages, groups, and even ads." As of writing, it's not known who is behind the campaign, although the threat actors have set up over 100 fake websites and malicious Chrome extensions. Google, for its part, has taken down the extensions. To mitigate risks, users are advised to stick with verified developers before downloading extensions, review requested permissions, scrutinize reviews, and refrain from using lookalike extensions. That said, it's also worth keeping in mind that ratings could be manipulated and artificially inflated by filtering negative user feedback. DomainTools, in an analysis published late last month, found evidence of extensions impersonating DeepSeek that redirected users providing low ratings (1-3 stars) to a private feedback form on the ai-chat-bot[.]pro domain, while sending those providing high ratings (4-5 stars) to the official Chrome Web Store review page. Found this article interesting? Follow us on Twitter  and LinkedIn to read more exclusive content we post. SHARE    

May 19, 2025Ravie LakshmananBrowser Security / Vulnerability Mozilla has released security updates to address two critical security flaws in its Firefox browser that could be potentially exploited to access sensitive data or achieve code execution. The vulnerabilities, both of which were exploited as a zero-day at Pwn2Own Berlin, are listed below - CVE-2025-4918 - An out-of-bounds access vulnerability when resolving Promise objects that could allow an attacker to perform read or write on a JavaScript Promise object CVE-2025-4919 - An out-of-bounds access vulnerability when optimizing linear sums that could allow an attacker to perform read or write on a JavaScript object by confusing array index sizes In other words, successful exploitation of either of the flaws could permit an adversary to achieve out-of-bounds read or write, which could then be abused to access otherwise sensitive information or result in memory corruption that could pave the way for code execution. The vulnerabilities affect the following versions of the Firefox browser - All versions of Firefox before 138.0.4 (including Firefox for Android) All versions of Firefox Extended Support Release (ESR) before 128.10.1 All versions of Firefox ESR before 115.23.1 Edouard Bochin and Tao Yan from Palo Alto Networks have been credited with finding and reporting CVE-2025-4918. The discovery of CVE-2025-4919 has been credited to Manfred Paul. It's worth noting that both shortcomings were demonstrated at the Pwn2Own Berlin hacking contest last week for which they were awarded $50,000 each. With web browsers continuing to be an attractive vector for malware delivery, users are advised to update their instances to the latest version to safeguard against potential threats. "Neither of the attacks managed to break out of our sandbox, which is required to gain control over the user's system," Mozilla said in a statement. "Despite the limited impact of these attacks, all users and administrators are advised to update Firefox as soon as possible." Found this article interesting? Follow us on Twitter  and LinkedIn to read more exclusive content we post. SHARE    