CySecurity News - Latest Information Security and Hacking Incidents

Beagle Backdoor Claude AI Malvertising malware User Privacy

Fake Claude AI Site Spreads New Beagle Windows Backdoor – Here’s How to Stay Safe

Cybercriminals have launched a sophisticated malvertising campaign using a fake Claude‑AI website that installs a new Windows backdoor called “Beagle,” highlighting how attackers are weaponizing the popularity of AI tools against software developers. The deceptive site, reachable through sponsored search results, mimics Anthropic’s legitimate Claude interface and lures users into downloading what appears to be a productivity‑oriented “Claude‑Pro Relay” tool but is in fact a poisoned installer.

Modus operandi

The malicious domain claude‑pro[.]com presents a stripped‑down clone of the official Claude design, using similar colors and fonts to create a veneer of legitimacy. However, most navigation links on the page simply redirect back to the homepage, and the only functional element is a large download button that serves a 505‑MB archive named Claude‑Pro‑windows‑x64.zip, which contains a trojanized MSI installer. Users who bypass standard security hygiene—such as verifying the URL or ignoring suspicious “sponsored” tags—end up deploying this bundle on their machines.

Once the MSI executes, it drops three files into the Windows Startup folder: NOVupdate.exe, NOVupdate.exe.dat, and a malicious DLL named avk.dll. The first file is a legitimate, digitally signed updater from G Data security software, which attackers abuse via DLL sideloading to load the malicious avk.dll instead of the genuine library. This DLL decrypts the encrypted data file, then executes the open‑source in‑memory loader DonutLoader, which in turn deploys the final payload—the Beagle backdoor—entirely in memory to evade disk‑based detection.

Beagle backdoor capabilities

Beagle is a lightweight but dangerous Windows backdoor that gives attackers remote control over an infected system. It supports a small set of commands such as running arbitrary shell commands, uploading and downloading files, creating and renaming directories, listing folder contents, and uninstalling itself to destroy evidence. The malware communicates with its command‑and‑control server at license[.]claude‑pro[.]com over TCP port 443 or UDP port 8080, encrypting traffic with a hardcoded AES key to make network monitoring more difficult.

Attribution and broader implications Security researchers have not yet pinned the campaign to a specific named threat group, but they note technical overlaps and suggest the same actors behind the PlugX malware family may be experimenting with this new payload. The fact that the attackers impersonate major security vendors in other related samples—such as Trellix, CrowdStrike, SentinelOne, and Microsoft Defender—points to a broader malvertising and supply‑chain‑style strategy.

How users and organizations can protect themselves

Organizations should block the domains claude‑pro[.]com and license[.]claude‑pro[.]com at the DNS and firewall level and search endpoints for NOVupdate.exe and avk.dll in Startup folders, which are strong indicators of compromise. End users, especially developers, must download Claude and similar AI tools only from verified official domains, treat sponsored search results with skepticism, and verify URLs before clicking installers. Updated endpoint protection, EDR logging, and user‑awareness training on AI‑related phishing and malvertising are critical to mitigating this evolving threat.

Meta Challenges Ofcom Over Online Safety Act Fees and Penalties



 

Meta Privacy Concerns

child online safety laws Cyber Security Digital Security Meta Meta Platforms Meta Privacy Concerns

Meta Challenges Ofcom Over Online Safety Act Fees and Penalties

Challenging new rules, Meta - owner of Facebook and Instagram - is taking Ofcom to the High Court amid disputes about charges tied to the Online Safety Act. The legal move stems from disagreements on how costs and fines are set by the UK's communications watchdog.

July 2025 marked the start of a legal shift meant to curb damaging material on internet services. Funding oversight duties now fall partly on big tech firms, each paying yearly charges based on global earnings. These payments support Ofcom’s work monitoring digital spaces. Rules took effect without delay once enacted. Revenue ties ensure contributions scale with company size. Later in 2025, new rules took effect targeting firms with annual earnings above £250 million.

These apply specifically to digital spaces like social networks and search tools - any platform allowing user-generated posts falls under scrutiny. While scale matters, the core focus remains on interactive online environments. Revenue size triggers obligation; activity type defines scope. What stands out is how Meta views the regulator's approach to setting operational charges and potential fines as skewed, placing too much burden on just a few major tech players. Shaped by courtroom arguments, legal representatives emphasized that today’s framework demands disproportionate contributions from firms like theirs.

Though the Online Safety Act applies across a wide range of online services, the cost structure reflects something narrower in practice. One outcome - seen clearly - is that even minor shifts in methodology could alter financial exposure significantly. Behind these figures lies an assumption: larger platforms must pay more simply because they can. Yet the law itself does not single them out for heavier obligations.

Instead, what emerges is a system where scale becomes a proxy for liability without clear justification. Disputing the method behind calculating eligible international income forms part of the legal argument. Court documents show Meta arguing penalties ought to reflect earnings only from UK-based operations, not total global turnover. Should firms fail to meet online safety duties, penalty amounts might reach 10% of global turnover - or £18 million - whichever figure exceeds the other.

Another layer emerges where Meta contests methods used to assign sanctions if several units within one corporate family share fault. Later in London, at an early court session, officials heard that Epic Games - creator of Fortnite - and the Computer and Communications Industry Association might ask to join the legal matter. The possibility emerged through statements presented to the High Court.

Later this year, more sessions will follow after Mr Justice Chamberlain pointed to matters of broad public significance in the case. Come October, a complete hearing should unfold. Following prior disputes over the Online Safety Act by various groups, litigation has now emerged again. Though distinct, last year’s challenge by the Wikimedia Foundation dealt with related rules on age checks - and ended in defeat.

Despite pushback, Ofcom stood by its method, saying fees and penalties followed directly from how the law is written. Rather than accept Meta's concerns, the authority insisted the system makes sure firms with major online influence support efforts to keep users safe. Still, Meta insists it will keep working alongside Ofcom, though parts of the rollout feel excessive to them. Even with their suggested adjustments, oversight bodies could still hand down penalties among the highest ever seen on British companies.

Crypto at Risk: Experts Believe Quantum Threat Arriving by 2030



 

Technology

AI Bitcoin Crypto ETH Etherium Quantum Computing Technology

Crypto at Risk: Experts Believe Quantum Threat Arriving by 2030

A recent report has warned that cryptographic foundations that secure trillions of dollars in digital currency can be hacked by quantum computers within the next four to seven years, and the blockchain industry is not prepared for damage control.

About quantum computing and threats

Project Eleven, a quantum security firm, published a report that said these quantum computers, even one, is powerful enough to hack the elliptic curve digital signatures securing Ethereum, Bitcoin, and other big blockchains. Experts say they won’t exist beyond 2033, and may end soon by 2030. The window for action is closing fast. According to the report, “Migration to quantum-resistant cryptography is no longer optional but imperative for any blockchain system expected to be trusted and secure into the future."

Why is quantum computing so fast?

Recent innovations have significantly lowered the hardware bar needed to launch such attacks. A breakthrough Google paper said that breaking the elliptic curve cryptography threshold could be achieved within 1,200 logical cubits, and less than 90 minutes of computing time on a supercomputing hardware.

Google has put a Q-Day (like D-day) at 2032. Project Eleven’s research has decreased the timeline by two years: 2030. The report estimates that 6.9 million Bitcoin (one third of the total estimated supply) have already been leaked on-chain, exposed to the potential quantum attack. For ETH, exposure is more, with over 65% of all ETH held in quantum-exposed addresses.

Why are blockchains weak against quantum computing?

The public ledgers and bearer-instruments offer no security. Blockchains has no scam department, no redressal platform for stolen funds, and no chargeback measures. If a quantum hacker recovers a private key and steals money, the loss is permanent. The transition problem is further fouled by slow-moving blockchain governance.

What makes blockchains particularly vulnerable, the report explains, is that their public ledgers and bearer-instrument design offer no safety net. Unlike a bank, a blockchain has no fraud department, no chargeback mechanism, and no way to reverse a forged transaction. Once a quantum attacker recovers a private key and drains a wallet, the loss is permanent.

Why is crypto migration difficult?

Bitcoin SegWit upgrade took more than two years to complete whereas ETH’s transition of proof stake took around 6 years to build. Quantum migration reaches the most basic layer of any blockchain mechanism.

The tech world has already started moving. More than half of web traffic (human) is currently post-quantum encrypted, Cloudflare data from December 2025 said.

Is the digital industry prepared?

The digital asset industry lacks preparedness. Crypto developers are suggesting various proposals but these plans will take years to execute while the threat is already brushing businesses and users.

"The internet has already moved," the report added. "The digital asset industry—which arguably has more at stake because blockchains directly protect bearer value with the exact cryptographic primitives that quantum computers threaten—has barely started."

GitHub Token Exposure at Grafana Triggered Codebase Theft Incident



 

Supply Chain Security

API Token Exposure Cloud Cyber Extortion Cyber Threats DevSecOps Security GitHub Token Leak Grafana Breach Supply Chain Security

GitHub Token Exposure at Grafana Triggered Codebase Theft Incident

Following the acquisition of a privileged GitHub token tied to Grafana Labs' development environment, a threat actor quickly escalated the initial credential exposure into a significant source code security incident. It was possible for the attacker to gain access to the company's private GitHub infrastructure, extract internal code repositories, and then attempt to extort payment from the organization via unauthorized access.

In addition to revoked credentials quickly, Gloria Labs launched an internal forensic investigation to determine the origin of the exposure and limit further risks. In spite of the fact that the breach resulted in access to sensitive development assets, the company announced that investigators found no evidence of data compromise, disruption of operations, or unauthorized access to user environments as a result of the breach.

Grafana’s widespread use in modern observability environments has drawn significant attention across the cybersecurity community due to the platform’s widespread role in monitoring infrastructure, cloud workloads, applications, and telemetry systems through centralized dashboards and analytics. The incident has attracted significant attention across the cybersecurity community.

In the course of the investigation, Grafana Labs disclosed that after detecting unauthorized activity, its security team initiated an immediate forensic response, eventually tracing the source of credential exposure and revoking the compromised access token in order to prevent further intrusion. Additionally, additional defensive controls were implemented across the company's development environment as part of its efforts to contain and harden the environment.

Afterwards, the threat actor attempted to extort the organization by requesting payment in exchange for delaying publication of the stolen data, according to the disclosure. Grafana, however, chose not to engage in ransom negotiations, aligning its response with Federal Bureau of Investigation guidance, which has consistently emphasized that paying extortion demands does not ensure data recovery nor prevent future misuse of stolen information.

A number of federal authorities have warned against ransom payments, stating that they rarely ensure suppression of stolen data and often contribute to additional criminal activity targeting technology providers and enterprise platforms.

The exact timeline of the attack or the length of time the attacker was permitted access to Grafana Labs' GitHub environment have not been disclosed, as only that the incident has recently been discovered. It is also noteworthy that the company did not explicitly attribute the intrusion to a specific threat actor.

However, various cyber threat intelligence reports, including Halcyon and Fortinet FortiGuard Labs assessments, have linked claims surrounding the incident with CoinbaseCartel, a collective of data extortionists. It has been noted that the group is an emerging extortion-focused operation that emerged in late 2025 and has operational overlap with criminal ecosystems such as ShinyHunters, Scattered Spider, and LAPSUS$ based on public statements released by Grafana.

According to the company's public statements, investigators believe that the intrusion occurred due to the compromise of privileged authentication tokens used in Grafana's development process. As a result, these tokens are frequently used to authenticate automated processes, integrations, and development workflows without requiring repeated manual logins. Although highly beneficial to operational efficiency, exposed tokens can also serve as high-value attack vectors when given broad permissions.

In this case, Grafana Labs' GitHub environment was compromised as a result of a compromised token that allowed the attacker access to private source code repositories within Grafana Labs. Despite the company's assertion that no customer information, user environments, or operational systems were compromised, the exposure of proprietary source code remains a significant security concern within software supply chain environments.

Although Grafana stated that customer environments were not affected, unauthorized access to proprietary source codes remains a serious concern, as attackers have the capability of analyzing internal architecture, configurations, or development logic to identify vulnerabilities that may later be used to conduct targeted attacks or other supply chain risks.

Grafana is widely deployed observability technology, and therefore the security of its development infrastructure is of particular importance. Attacks against software vendors may result in downstream risks affecting customers, cloud deployments, as well as broader enterprise environments linked by modern DevOps and observability pipelines. Upon tracking the threat intelligence associated with the incident, it has been determined that the operators behind the claimed attack are primarily engaged in data theft and extortion operations rather than conventional ransomware operations that encrypt files.

Over 170 victims have been linked to the group across sectors such as healthcare, transportation, manufacturing, and technology, reflecting the growing trend toward cyber-attacks that focus on data theft and extortion. There has been no public announcement by Grafana Labs regarding which repositories or internal projects were accessed during the breach, indicating that there is no clear understanding of the scope of the material that was downloaded. Grafana Labs has not disclosed which repositories were accessed during the breach.

In addition to Grafana Cloud, Grafana's managed cloud monitoring platform is widely used across enterprise environments for observing observability. In addition to the disclosure, cyber attacks aimed at extortionating software vendors and cloud service providers are also becoming increasingly aggressive. Following threats of leaking large volumes of data supposedly associated with schools and universities across the United States, Instructure reportedly agreed to negotiate with threat actors connected to ShinyHunters following an alleged agreement to negotiate.

Grafana Labs' decision to reject the extortion demand reflects a growing industry debate concerning ransomware economics, incident response strategies, and the long-term consequences of compensating cybercriminals. A company statement in accordance with advice issued by the Federal Bureau of Investigation stated that paying attackers would not guarantee the suppression of the stolen material nor eliminate the possibility of future abuse, resale, or repeated extortion attempts.

The company notes that organizations have no assurance that the stolen information will actually be removed after payment, which makes ransom negotiations risky and uncertain from an operational perspective. The incident emphasizes the high value of authentication tokens, API credentials, and machine-level secrets within enterprise environments, in addition to the breach itself.

In order to reduce the risk of token-based intrusions and software supply chain attacks, security teams are increasingly recommending implementing measures such as short-lived credentials, least privilege access, credential rotation, and multi-factor authentication. They also recommend continuous monitoring of repositories and continuous delivery pipelines.

The enterprise attack surface has been increasingly centered around GitHub repositories, package distribution systems, internal build pipelines, and cloud-based engineering environments, which require security controls comparable to those protecting production infrastructure. Grafana Labs has gained attention for its relatively transparent disclosure approach despite the seriousness of the intrusion.

A statement from the company outlined the compromise, clarified what investigators believe remains unaffected, disclosed the attempted extortion component, and indicated that further details may become apparent as the forensic investigation proceeds. At present, the known impact appears to be limited to unauthorised access and download of internal source code repositories, with no evidence suggesting that customer environments, operational systems, or personal information has been compromised.

Grafana remains closely monitored across the cybersecurity community, as it is widely used throughout production observability stacks and cloud-native enterprise environments around the world. Despite Grafana Labs' assurance that customer systems and personal data were not affected, the incident highlights the increasing importance of securing development infrastructure, access credentials, and cloud-connected engineering environments against increasing sophistication in extortion-focused threats.

Canadian Privacy Regulators Say OpenAI Violated Federal and Provincial Privacy Laws



 

Data Privacy Laws

AI Chatbots AI Model AI Privacy Canada ChatGPT Privacy ChatGPT. OpenAI Cyber Privacy Cyber Security Data Privacy Laws

Canadian Privacy Regulators Say OpenAI Violated Federal and Provincial Privacy Laws

After months of scrutiny, Canadian oversight bodies determined OpenAI did not meet several national and regional data protection standards while developing its AI systems. This outcome emerged from a coordinated review spearheaded by federal Privacy Commissioner Philippe Dufresne, working together with counterparts in Alberta, Quebec, and British Columbia.

What stood out in the findings was a pattern of data handling at OpenAI - massive volumes of personal details gathered, yet lacking strong protections or clear approval from affected people. Because of this approach, authorities concluded it clashed with rules set by Canada’s privacy law, known formally as PIPEDA, guiding how firms manage private data while conducting commercial activities.

The way ChatGPT and similar artificial intelligence models were developed raised notable questions for oversight bodies. A key point centered on data collection practices - information about people pulled from open internet resources and external databases, often without clear notice to those affected. Officials pointed out that many users remain unaware their details might feed into machine learning processes.

Another concern emerged around control: few practical options let individuals inspect, update, or request deletion of their data linked to these systems’ training records or responses. Oversight groups stressed that current safeguards fall short in offering real transparency or user agency. Questions arose about how dependable ChatGPT's answers really are.

Some pointed out that current methods for managing false or confusing replies fall short - especially if private information is at stake. Even so, Canadian privacy authorities observed OpenAI engaging throughout the probe, committing in advance to adjustments meant to bring operations into line with national data rules. Following these steps, it appears older versions of the AI were phased out due to shortfalls in compliance, while new filters emerged - meant to spot and obscure details like contact numbers or full names across both open-access and legally obtained training collections.

Some time soon, OpenAI will adjust how it explains the role of user chats in training its systems. A new phase involves more noticeable alerts for people using ChatGPT without logging in. These notices aim to guide visitors away from submitting private details. How exchanges help shape upcoming models will also become part of that message. Updates are meant to surface key points earlier in the experience.

Further changes include streamlining how users access their data, while offering straightforward steps for disputing AI-generated inaccuracies. Officials emphasized protections for young relatives of well-known individuals - models must now avoid revealing personal details like names or birthdays if the child is not publicly recognized.

Later scrutiny emerged when news surfaced connecting OpenAI to alarms tied to a violent event in Tumbler Ridge during early 2026, reigniting interest in an inquiry first begun in 2023. Though internal signals about the individual's activity were reportedly noticed earlier, officials claimed the firm failed to forward such red flags to Canadian authorities. Because of what followed, oversight bodies emphasized better coordination among artificial intelligence developers, police units, and public health offices whenever physical harm appears likely.

Rather than wait, expectations now lean toward faster information sharing across these groups. Pressure mounts globally as scrutiny increases on firms using artificial intelligence, pushing them toward stronger safeguards for personal data. How information is gathered and applied in training powerful models now faces closer examination.

Greater openness about methods has become harder to avoid. Responsibility for outcomes ties directly to practices behind massive data processing. Standards shift under persistent demands for clearer conduct.

Delivery-Label Scam: How Amazon & Flipkart Boxes Can Steal Your Data



 

Shipping Labels Identity Theft

Cyber Fraud Delivery Scam Parcel Security Data Privacy Shipping Labels Identity Theft

Delivery-Label Scam: How Amazon & Flipkart Boxes Can Steal Your Data

Scammers are exploiting discarded delivery boxes from Amazon and Flipkart to harvest personal information and launch sophisticated phishing attacks, so shoppers need to treat packaging as sensitive data rather than trash. Labels on parcels often include names, phone numbers, addresses and sometimes order details, which fraudsters collect from bins or common disposal areas and then use to make their outreach appear legitimate.

The attack begins with a simple, low-tech step: gathering boxes with intact shipping labels. Criminals extract the printed information and then contact the recipient posing as customer-care or rewards agents, leveraging the accurate personal details to build trust quickly. Because the caller or message can reference the victim’s real name, number, and recent purchase, targets are more likely to engage and follow instructions, which typically include clicking a link or sharing an OTP to “confirm” a cashback, refund, or prize.

Clicking the supplied link or following caller instructions is where the compromise happens. Victims are often led to phishing pages or prompted to install malicious apps that capture credentials, banking details, and OTPs, or to enter payment information directly on fake forms; these steps can lead to immediate financial loss and longer-term account takeover. Scammers sometimes combine this with social-engineering scripts—urgent tones, limited-time offers, or threats of cancelled orders—to pressure victims into acting without verification.

Safety tips

Protection is straightforward but requires habit change. Before discarding any parcel, remove or destroy the shipping label—tear it off, shred it, black it out with a permanent marker, or use an identity-protection roller stamp to obscure personal data. Never click links or install apps sent by unknown numbers, and verify any unexpected offers or refund requests directly through the official Amazon or Flipkart apps or websites rather than through messages or calls. Treat unsolicited calls that reference personal order details with skepticism, and never share OTPs, passwords, or bank information even when the caller appears informed.

This scam becomes especially active around big sale events when large volumes of deliveries increase the supply of labelled packaging and scammers’ opportunities to find usable targets. A few seconds spent removing labels and a little caution with links and calls can block an easy avenue criminals use to convert harmless cardboard into a source of identity theft and financial fraud.

4 Key Areas in 2026 for Organisation Safety Against Advanced AI Threats



 

Technology

agentic AI AI ChatGPT Cloud GenAI Internet Technology

4 Key Areas in 2026 for Organisation Safety Against Advanced AI Threats

2026 has not been a kind year to cybersecurity, as organizations and industries globally have been hit by ruthless cyberattacks.

2026 and cybersecurity

Cybersecurity entered 2026 under stress to deploy AI tech while building foundations for a quantum future. Cybersecurity experts have to defend against advanced AI and hybrid attacks while facing talent scarcity, a rapidly shifting threat scenario, and rising operational challenges.

It is the first time that hackers have access to the same advanced enterprise-level tech that security experts are using to defend their digital assets.

Is the convergence good or bad?

Organizations are in need of the transformational advantage that Quantum computing promises, however, it also risks affecting the cryptographic infrastructure that protects today’s digital world. Worse, cyber attackers are getting together and outbeating experts.

Like experts, threat actors don’t mind playing the long game either, they gain initial access and stay hidden inside systems for longer periods of time. When the right opportunity arrives, they move laterally and hack important data that can affect operations, cause financial damage, and tarnish reputations.

So, what are these four key areas that businesses and users need to address or stay safe from?

1. System and skills problem

As per the ICS2 2025 report, 69% respondents suffered multiple cybersecurity breaches due to skill gaps. This is due to various factors such as budget constraints, misalignment in academia, and high enterprise demand.

2. Bug management shift to active exposure reduction

Hackers use GenAI to advance their attacks, scaling, and escape security experts. This reactive cycle delays response times, and gives just basic protection. What businesses need today is Continuous Threat Exposure Management (CTEM) approach that offers real-time visibility before flaws can be exploited. But the success depends on AI-based risk prioritization.

3. Advanced deepfake protection is the need of the hour

Reliability is the new attack vector. Deepfakes have plagued every digital aspect of human life. Traditional measures fail to address content due to AI, therefore AI-based protection is needed. Adaptive deepfake systems can address identity workflows and respond immediately to threats, flagging malicious activity and capturing attacks with detailed metadata for research and audit work.

4. Post-quantum protection

Quantum computing is making strides in applicability; if sufficiently advanced, the systems can break public-key cryptographic systems in ransomware attacks such as RSA, where hackers extort millions. Hackers are already using the “harvest now, decrypt later” approach, stealing coded data with no promise of returning it.

Thus, the National Institute of Standards and Technology (NIST) have advised to adopt post-quantum cryptography (PQC) and tracking quantum-vulnerable assets.

Trojanized DAEMON Tools Used to Deploy Persistent Backdoor Malware



 

Supply Chain Attack

Backdoor Malware Cyber Attacks cybersecurity threat DAEMON Tools information stealer Malware Campaign PowerShell Attack QUIC RAT Software Compromise Supply Chain Attack

Trojanized DAEMON Tools Used to Deploy Persistent Backdoor Malware

An innocent routine software update mechanism has been weaponized by attackers in order to distribute malware through official distribution channels, enabling a stealthy global supply-chain compromise. AVB Disc Soft authenticated digital certificates were used to sign trojanized builds as part of the operation that remained undetected for nearly a month.

By bypassing conventional trust and endpoint security mechanisms, these malicious packages were able to avoid triggering immediate suspicion. Kaspersky discovered that the campaign began on April 8, 2026, and resulted in thousands of infections in over 100 countries before the breach was detected on May 1, 2026.

Almost all infections were characterized by reconnaissance malware intended to gather system intelligence and establish persistence. However, a comparatively small number of carefully selected victims received advanced second-stage backdoors, suggesting a targeted attack on Russian, Belarusian, and Thai organizations involved in government, science, retail, and manufacturing.

Multiple core components of DAEMON Tools were modified, including DTHelper.exe, DiscSoftBusServiceLite.exe, and DTShellHlp.exe, and malicious functionality was embedded in versions 12.5.0.2421 through 12.5.0.2434, ensuring that execution occurs at startup while maintaining the appearance of legitimate software functionality.

According to the forensic analysis, the attackers had embedded their malicious framework within several trusted DAEMON Tools binaries, including the DTHelper.exe, DiscSoftBusServiceLite.exe, and DTShellHlp.exe that can be found within the installation directory of the application. Because the compromised binaries were signed by authentic AVB Disc Soft signing certificates, operating systems and endpoint security products perceived the compromised binaries as trustworthy, reducing the probability of immediate detection.

It has been determined that every time the affected binaries are executed during system startup, the CRT initialization routine initiates hidden backdoor functionality, initiating a dedicated background thread aimed at quietly establishing outbound communication with attacker-controlled infrastructure during system startup.

Throughout the attack, the malware repeatedly sent HTTP GET requests to a typosquatted domain that closely mimicked the legitimate DAEMON Tools download portal, as a method of mixing malicious traffic with expected software communications. According to WHOIS records, the fraudulent domain was registered on March 27, approximately one week before the supply chain intrusion occurred, indicating deliberate preparation of infrastructure prior to the attack by the campaign's operators.

Based on an analysis of the command-and-control infrastructure, it appeared that compromised systems were able to receive remotely issued shell commands via cmd.exe and PowerShell, which would allow attackers to download and execute additional payloads dynamically.

PowerShell's WebClient functionality was utilized to retrieve executable files from an Internet server located at 38.180.107[.]76 before silently executing them from temporary system directories and deleting all traces afterwards. In the course of the investigation, envchk.exe, a .NET-based information collector that researchers determined was intended to perform extensive reconnaissance on infected machines, was identified as one of the primary secondary payloads.

In the malware's source code, embedded Chinese-language strings suggest that the malware's operators are probably Chinese-speaking, but no official affiliation has yet been established for the threat group. This reconnaissance utility collected a broad range of information regarding the host, including MAC addresses, hostnames, DNS domains, installed software inventories, running process lists, system locale configurations, and other host information.

Following data collection, the collected data is transmitted back to attacker-controlled infrastructure via structured HTTP POST requests, providing the operators with a detailed profile of the compromised environment before deciding whether to escalate the intrusion. Unsuspecting users were infected when they downloaded and installed trojanized yet legitimately signed installers for DAEMON Tools, which executed malicious code contained within trusted application components without the user knowing it.

After activation, the implanted payload established persistence mechanisms intended to survive reboots, as well as enabled the installation of a covert backdoor capable of communicating with remote attackers when the system is started.

The command infrastructure was also capable of dynamically delivering additional malware stages based on the victim’s profile and operational significance. It is generally considered to have functioned as a reconnaissance-oriented information stealer tasked with gathering system identifiers, including hostnames, MAC addresses, running processes, installed applications, and locale configurations, before transmitting the harvested telemetry to the operators for the purpose of assessing the environment and prioritizing victims.

The first-stage profiling phase of the investigation resulted in an evaluation of selected systems for further compromise. Using a lightweight backdoor that is capable of executing arbitrary commands, downloading files, and running malicious code directly in memory, selected systems were escalated to a second-stage compromise.

The attack on a Russian educational institution was escalated by the attackers by using QUIC RAT, a remote access malware strain capable of supporting a variety of communication protocols, as well as injecting malicious code into legitimate processes so that they could operate stealthily after the compromise.

Despite utilizing software distributed through official channels, the DAEMON Tools breach remained undetected for nearly a month as a highly coordinated and technically mature supply-chain intrusion. An investigation into DAEMON Tools installations conducted on or after April 8 was advised to conduct extensive threat-hunting operations to monitor for abnormal system behavior and unauthorized network activity related to the compromise period.

Researchers have avoided formally identifying the threat actor behind the campaign, but linguistic artifacts embedded within its first stage strongly suggest that Chinese-speaking operators were responsible. Following earlier compromises involving eScan, Notepad++, and CPU-Z, the incident also illustrates the rising trend of software supply-chain attacks throughout 2026. In parallel with these campaigns, the increasing importance of trusted software ecosystems becoming high-value attack surfaces for sophisticated threat groups continues to be demonstrated, including Trivy, Checkmarx, and Glassworm, which target software repositories, development packages, and browser extensions.

The DAEMON Tools compromise proves that modern supply-chain attacks are not limited to niche targets or underground software ecosystems, but are increasingly exploiting widely used consumer and enterprise applications. The attackers developed their attack strategy by leveraging trusted software certificates and official distribution channels in order to disguise malicious activity as legitimate software behavior while quietly gaining access to potentially high-value environments across multiple countries.

Security researchers have concluded that organizations must evolve beyond traditional trust-based security models and embrace continuous monitoring, behavioral detection, and software integrity validation practices that will enable them to identify malicious activity, even within applications that appear legitimate and have been signed. A contemporary supply-chain intrusion illustrates how a single compromised software update can quickly escalate into a global cyber risk with far-reaching operational and national security consequences.

Election Commission Says ECINET Withstood Over 68 Lakh Cyberattack Attempts During Poll Counting



 

Election Security

Cyber Attacks Digital Infrastructure ECI Election Security

Election Commission Says ECINET Withstood Over 68 Lakh Cyberattack Attempts During Poll Counting

The Election Commission of India (ECI) said its digital election infrastructure faced more than 68 lakh malicious online hits on the day votes were counted for the recently concluded Assembly elections, with attempts originating from both domestic and overseas sources. According to election officials, the attacks targeted several online systems operated by the Commission, including the public election results portal, but were contained using existing cybersecurity protections.

Officials stated that despite the unusually high volume of hostile traffic, there was no disruption to counting operations or public access to election-related services.

The attacks were directed at ECINET, the Commission’s integrated election management platform that now combines over 40 separate election applications and digital portals into a unified system. The platform is used to manage multiple election-related functions, including monitoring, reporting, voter services, and administrative coordination.

On counting day, May 4, ECINET reportedly processed an average of nearly 3 crore hits every minute. Across all polling phases conducted on April 9, 23, and 29, the platform recorded a total traffic load of 98.3 crore hits, reflecting the scale at which India’s election infrastructure now operates digitally.

The Commission officially launched ECINET in January 2026 after testing its beta version during the Bihar Assembly elections in November 2025. Since then, the application has crossed 10 crore downloads, indicating rapid adoption among election officials, staff, and users accessing poll-related information and services.

Election authorities said the platform played a major operational role during the elections across five states and Union Territories, along with bypolls conducted during the same period. According to officials, ECINET enabled real-time monitoring of election activities, accelerated reporting processes, and improved administrative coordination between different election units. Authorities also said the centralized system helped increase transparency by reducing delays in communication and data sharing.

Cybersecurity analysts have repeatedly warned that election infrastructure has become an increasingly attractive target for malicious cyber activity because such systems process large amounts of real-time public information under intense public scrutiny. During counting periods, election portals often experience massive spikes in traffic as citizens, media organizations, and political workers continuously refresh result dashboards. Security researchers note that these high-traffic periods can also create opportunities for malicious actors to disguise harmful requests within normal user activity.

While the Election Commission did not disclose the technical nature of the 68 lakh malicious hits, such traffic typically includes automated bot requests, denial-of-service attempts, malicious scanning activity, or repeated unauthorized access attempts aimed at slowing systems or overwhelming servers.

The Commission also introduced a new QR code-based photo identity verification system for counting centres during the election process. On counting day alone, more than 3.2 lakh QR codes were generated through ECINET to regulate entry into counting venues. Officials said the system was introduced to ensure that only authorized personnel could enter restricted areas, reducing the possibility of unauthorized access at highly sensitive counting locations.

According to the Commission, this was the first time the QR-based access system had been deployed across all five states and Union Territories simultaneously. The ECI has now decided to adopt the system as a standard security measure for future Lok Sabha and state Assembly elections.

The increasing dependence on centralized digital infrastructure has pushed election management beyond traditional ballot security into the broader domain of cybersecurity, network resilience, identity verification, and real-time system monitoring. As more election operations move onto integrated digital platforms, experts say continuous monitoring and infrastructure hardening will become essential to maintaining uninterrupted electoral processes at national scale.

Quasar Linux Malware Targets Developers in Stealthy Supply Chain Attack



 

quasar

chain attacks Cybersecurity Digital Supply Chain Risk Linux malware quasar

Quasar Linux Malware Targets Developers in Stealthy Supply Chain Attack

A newly discovered Linux implant called Quasar Linux, or QLNX, is a serious threat because it goes after the people and systems that build software. Instead of behaving like ordinary malware, it is designed to quietly take root in developer and DevOps environments, steal valuable credentials, and open the door to supply-chain attacks.

QLNX is dangerous because it combines several attack techniques in one package. Trend Micro says it can function as a rootkit, a backdoor, and a credential stealer, while also running filelessly, wiping logs, spoofing process names, and removing its original binary from disk to make investigation harder. It also uses multiple persistence methods, including LD_PRELOAD, systemd, crontab, init.d scripts, XDG autostart, and .bashrc injection, so it can keep coming back even if part of it is removed.

The malware’s main prize is access to developer secrets. Researchers say it targets credentials tied to npm, PyPI, GitHub, AWS, Docker, Kubernetes, Terraform, and other tools that are deeply embedded in modern software delivery pipelines. If attackers get those tokens or keys, they can publish malicious packages, tamper with builds, or move from one system into cloud infrastructure and CI/CD environments.

What makes the threat especially troubling is how stealthy it is. Trend Micro found that QLNX can dynamically compile rootkit and PAM backdoor components on the victim host using gcc, which helps it blend in with normal Linux activity. It also harvests clipboard contents, SSH keys, browser profiles, and authentication data, giving attackers a wide view into how developers work and where their secrets are stored.

The broader issue is that developer machines have become high-value targets in the software supply chain. One compromised workstation can expose publishing pipelines, cloud accounts, and internal codebases, so the impact may spread far beyond the original victim. The safest response is to treat developer endpoints like crown-jewel systems: monitor for unusual persistence, restrict secret storage, rotate tokens quickly, and assume a stolen workstation could become the first step in a wider breach.

Ubuntu DDoS Attack Disrupts Installs Updates and Canonical Infrastructure



 

Infrastructure Security

API Cyber Attacks Cybersecurity Attack DDoS DDOS Attacks Hacker attack Infrastructure Security

Ubuntu DDoS Attack Disrupts Installs Updates and Canonical Infrastructure

A wave of traffic overwhelmed systems, briefly halting downloads, patches, and web resources managed by Canonical - the team responsible for Ubuntu Linux. Outages stretched nearly twenty-four hours, blocking access to essential tools during the incident.

Midway through the disruption, Canonical confirmed issues affecting its online systems, calling them a prolonged international cyber incident. With efforts already underway to bring functions back online, progress reports were expected later via verified sources after conditions improved.

Not just external sites felt the impact - insights from casual chats on unaffiliated Ubuntu message boards pointed to deeper issues. Failures popped up across several core functions: the security API stumbled, repository access broke, setup tools froze, package upgrades failed. When the outage struck, countless machines could neither pull patches nor start clean installs. The ripple spread wider than first assumed.

A claim of responsibility emerged afterward, attributed to an entity calling itself The Islamic Cyber Resistance in Iraq 313 Team. Supposed messages circulated on Telegram suggest they relied on a service named Beemed - one that facilitates distributed denial-of-service attacks - to execute the incident. While details remain sparse, the method points toward accessible cyber tools being leveraged for disruptive purposes. Heavy network floods emerge when tools like Beamed hand out DDoS power to anyone willing to pay, masking harm behind so-called "testing" labels.

Instead of building safeguards, some misuse these setups to drown web systems in endless data streams. With advertised force climbing toward 3.5 terabits each second, one sees how readily extreme digital pressure becomes a purchasable option. A single flood of fake signals can overwhelm digital infrastructure when launched from countless hijacked gadgets online.

Such an event forces critical systems to choke on excessive demand, blocking normal access. Real people experience delays or complete service failures as their requests get lost in chaos. Machines turned into unwilling helpers generate relentless noise instead of useful responses. Performance drops sharply once capacity limits are breached without warning. Genuine interactions fade under pressure from artificial congestion.

Most times, hacking groups start by slipping malicious software onto gadgets, sometimes using poor login codes instead of strong ones. From there, machines already taken over get bundled together - forming massive clusters run from far away via command centers online. These hijacked setups often change hands in hidden digital bazaars; launching short outages becomes possible for cheap, while heavier assaults require deeper spending.

What follows? Buyers pick time-limited chaos or go all-in for longer surges. Surprisingly, more DDoS attacks happen now due to widespread access to self-running malware that exploits weak device protections across countries. While strong networks may resist some threats, major companies still face interruptions since hackers pair huge bot-driven data floods with focused attack plans.

The Ubuntu event underscores how fragile key open-source tools have become - tools that developers, businesses, and public agencies depend on worldwide. When update servers or security interfaces go offline briefly, ripple effects follow. Patching halts. System rollouts stall. All of this unfolds while digital attacks are already underway.

High Court Squashes Ban for Sim-Swap Fraud, Says Zero Customer Liability



 

Tech

Banks Cyber Security Data Data Leak digital scam identity Scam Tech

High Court Squashes Ban for Sim-Swap Fraud, Says Zero Customer Liability

In an important ruling amid surging digital financial fraud attacks, the Bombay HC sided with the customer protection norms. It directed Bank of Baroda to return Rs. 1.24 crore to the victim private firm that lost money in a SIM-swap case. The court stressed that if a consumer reports fraud promptly in time, “zero liability” is ruled, and the bank must reimburse the losses.

Private company reported the incident immediately

The order was given by a division bench of the HC, which included Justices Manjusha Deshpande and Bharati Dangre, when private company PNP Polytex (based in Mumbai) submitted a petition. Polytex alleged that Rs.1.24 crore had been stolen from its bank accounts illegally and without knowledge.

About court proceedings

As per the submissions to the court, the firm informed the bank soon after finding malicious transactions and asked the accounts to be frozen. The bank could only save Rs. 47.8 lakh, the remaining money was already stolen by the hackers. After this, the firm moved to HC for help.

Later, enquiry revealed that the scam was done using a SIM-swap tactic, where hackers get control of the target’s registered contact number. This lets the hackers intercept OTPs and do banking transactions without the account owner's consent and knowledge. The high court found that the scam was done by third-parties, and showed no evidence of negligence on consumer’s end.

What is RBI’s zero liability rule?

During the proceedings, the court referred to the July 6, 2017 statement given by the RBI, which laid down the customer protection guidelines in incidents of illegal electronic banking transactions. According to the circular, the consumers are entitled to zero liability if they report fraud transactions within 72 hours (three days).

In the judgement, the high court stressed that if a customer informs the bank about a scam or fraud, it is the duty of the bank to return the disputed amount back to the victim’s account. The court also said that the burden of proving customer negligence is on the bank too.

The court rejected the bank's defenses that it had followed the due process and security measures, and the bench labelled the argument as a “lame excuse,” saying that such mechanisms become powerless when a SIM card is hacked. The court also attributed another ruling in an incident where HDFC bank was held liable under similar situations.

Bank will return stolen amount with interest

After revising the previously frozen funds, the High Court ordered the bank to return the remaining sum plus 6% interest within eight weeks.

Apple Account Data and Bluetooth Signals Tie Suspect to Crypto Robbery



 

Siri Vulnerability

Apple Security Bluetooth Forensics Crypto Wallet cryptocurrency theft Cybercrime Investigation digital surveillance FakeWallet Campaign iOS security malware Siri Vulnerability

Apple Account Data and Bluetooth Signals Tie Suspect to Crypto Robbery

The App Store ecosystem has been infiltrated by a coordinated wave of fraudulent cryptocurrency wallet applications that exploit regional platform restrictions and user trust to steal credentials from iOS users. More than two dozen malicious apps have been identified as related to a campaign called "FakeWallet," which has been active since at least late 2025 and was designed to harvest passwords and private keys from unsuspecting users via the use of various malware programs.

During the early months of March, counterfeit wallet applications became prominent in search results within China’s App Store after they began appearing prominently in search results, posing a threat to the legitimacy of several legitimate crypto wallet services due to regulatory restrictions.

In addition to replicating the trusted wallet branding, abusing typosquatting techniques and embedding deceptive prompts leading users towards unofficial wallet downloads, the campaign blurred the distinction between genuine financial tools and malicious software, significantly increasing iPhone users' chances of committing cryptocurrency theft.

During technical analysis, Kaspersky determined that phishing applications were primarily used as delivery mechanisms for trojanized cryptocurrency wallet software to be installed via browsers. According to the researchers, malicious payloads are commonly embedded through third-party libraries embedded within the applications, despite several samples demonstrating direct modifications of the wallet code itself, indicating a more sophisticated level of tampering.

Through reverse engineering, special routines have been found that can intercept and exfiltrate recovery phrases as well as seed phrases, while simultaneously manipulating the wallet restoration process for recovering hot wallets. The investigation also identified two separate implants targeting cold wallets hosted on Ledger, extending the campaign's scope beyond software-based assets to hardware wallet users as well.

A counterfeit website impersonating Ledger's official platform was also discovered by researchers, which distributed malicious iOS application links and compromised Android wallet packages hosted on Chinese-language phishing websites outside of Google Play. It is unclear whether the malware modules had geographic enforcement mechanisms despite the infrastructure and linguistic indicators suggesting that Chinese-speaking victims were targeted.

It is of concern that the campaign may easily be extended to international targets based on some phishing prompts that dynamically adapt to the language settings of the infected application. Furthermore, the operation has been linked to the previously identified SparkKitty malware cluster, which was discovered last year, based on overlapping distribution tactics, cryptocurrency-centered targeting patterns, Chinese-language debugging strings within the malicious code, and the inclusion of SparkKitty-related components within several analyzed programs.

When the findings were disclosed to Apple, they were notified and the identified malicious applications have since been removed from the App Store. According to court records reviewed by Forbes, the incident occurred as a result of a targeted home invasion last month in Winnetka, where attackers allegedly used social engineering tactics to gain physical access to the victim's property.

Investigators reported that a man impersonating a food delivery driver approached the residence and knocked on the front door before at least four armed accomplices gained access moments after the resident responded. Once inside, the group demanded access to a secure safe as well as credentials related to online cryptocurrency accounts, emphasizing the increasing convergence between the targeting of digital assets and conventional violent crimes.

A report by authorities indicates that the operation failed in achieving its intended objective after the victim escaped the residence, leading the suspects to depart the scene without obtaining any known cryptocurrency assets.

In spite of the attempted robbery, organized groups have increasingly combined physical coercion with identity deception and intelligence-driven targeting to compromise high-value cryptocurrency holders. It is believed that the investigation developed into a broader criminal case involving Chicago rapper Lil Zay Osama, formally known as Isaiah Dukes, along with five additional suspects, were alleged to have kidnapped children and committed a violent cryptocurrency-related robbery.

Dukes has entered a not guilty plea to the latest charges after previously serving a 14-month prison sentence for unlawful possession of a machine gun in 2024. According to reports, investigators used unconventional but highly effective digital forensics methods in order to identify members of the group after one suspect connected his iPhone to a stolen getaway vehicle's Bluetooth interface.

The combination of the infotainment pairing logs and the subpoenaed Apple records provided authorities with information that allowed them to locate the connected device in a iCloud account belonging to Tyrese Fenton-Watson. The discovery was significant as it demonstrated how telemetry generated by connected consumer technologies, such as smartphone synchronization and in-vehicle wireless systems, is becoming an increasingly important tool for criminal investigations in modern times.

Technology and cybersecurity landscapes were also subject to increasing scrutiny due to the emergence of artificial intelligence, surveillance practices, and digital governance concerns. Anthropic's reported intention to broaden access to its advanced "Mythos" model, which was originally restricted to approximately 40 organizations due to concerns surrounding misuse of the system and offensive security applications. This model is designed with large-scale cyber vulnerability discovery capabilities and is designed to detect cyber vulnerabilities on a large scale.

Reports in The Wall Street Journal indicated that the company hoped to expand its availability to approximately 120 companies, though White House officials expressed reservations about both national security implications and the potential strain on Anthropic's infrastructure and disruption of government access to the technology that could result from excessive external usage.

In addition, further revelations indicated that the boundary between the deployment of AI, the privacy of users, and digital surveillance is increasingly blurred. In a report published by Wired, it was reported that the DHS had requested location and identification information from Google regarding a Canadian user who criticized the Trump administration, but it is unclear whether Google complied with this request.

Additionally, Meta disclosed that Facebook and Instagram were using artificial intelligence-driven bone structure analysis to detect whether users are under the age of 13. According to security researcher Jeremiah Fowler, nearly 90,000 screenshots allegedly extracted from a celebrity's smartphone had been exposed as a result of spyware exposure, including sensitive photos, financial records, and private conversations, further illustrating the degree of personal data risks associated with commercial surveillance tools.

A significant amount of industry attention was also drawn to Forbes' publication of its eighth annual AI 50 ranking in partnership with Mayfield, highlighting some of the leading private AI firms, including Harvey and ElevenLabs, along with emerging startups, including Gamma, Chai Discovery, and Rogo. In addition, the AI 50 Brink list highlighted early-stage companies that were expected to compete effectively with more established companies.

During the investigation, law enforcement agencies also recorded a notable operational success after cooperating with Meta and international authorities to dismantle nine cryptocurrency scam centers and arrest more than 275 individuals allegedly involved in fraudulent schemes targeting Americans. This marks a rare instance of coordinated action between the Department of Justice and China's Ministry of Public Security.

A report alleging that workers employed by contractor Sama encountered explicit and sensitive footage while annotating video captured through Ray-Ban smart glasses prompted Meta to be subjected to renewed scrutiny for its privacy oversight. As a result of these allegations, Meta terminated its relationship with Sama shortly before terminating its agreement due to an unmet standard, a claim Sama denied publicly.

Following the latest developments, the company issued a series of critical software updates to resolve vulnerabilities affecting Siri, the company's voice-based digital assistant, resulting in the potential for unauthorized access to sensitive user information on locked mobile devices. These updates further renewed attention to mobile device security. It was found that the assistant was capable of processing certain voice interactions even while the device was locked, allowing attackers who possessed iPhones or other Apple hardware to access contact information and additional private data without complete authentication if they had physical possession of the devices.

As a result, Apple introduced security enhancements as a means of limiting Siri's functionality when devices are immobilized. By doing so, Apple reduces the likelihood that unauthorized commands may be executed while the device is immobilized as well as strengthening protections against physical access attacks. Several products within Apple's ecosystem, including iPhone, Apple Watch, iPadOS, and macOS Ventura systems, have been patched as part of broader platform security updates to mitigate the vulnerabilities.

Several software updates have been recommended to ensure that vulnerabilities are fully mitigated across all supported devices, including iOS 17.6 and iPadOS 17.6, by using the standard settings, general, and software update process.

Collectively, these incidents reflect a rapidly evolving threat environment in which cybercrime, artificial intelligence, connected consumer technologies, and digital surveillance are becoming increasingly interconnected. This collection of cases illustrates how both attackers and law enforcement are leveraging the expanding data footprint created by modern devices and online services in order to infiltrate trusted app ecosystems with malicious cryptocurrency wallet campaigns as well as investigators using Bluetooth telemetry and cloud account records to investigate violent crimes.

Furthermore, growing concerns surrounding the discovery of vulnerabilities using artificial intelligence, spyware-linked data exposure, biometric analysis, and voice assistant security continue to increase pressure for technology companies to strengthen platform security measures while maintaining a balance between privacy, accessibility, and operational transparency.

Increasing sophistication and technical integration of cyber-enabled financial crime underscores the importance of proactive security updates, stricter application vetting, and enhanced awareness of consumers in increasingly interconnected digital ecosystems as cyber-enabled financial crime becomes more sophisticated and technologically integrated.

Chrome Quietly Installs 4GB AI Model on Users’ Devices Without Permission



 

Technology

AI model download Chrome privacy concerns ePrivacy Directive EU privacy laws Gemini Nano Google Chrome Technology

Chrome Quietly Installs 4GB AI Model on Users’ Devices Without Permission

lGoogle Chrome has reportedly begun silently downloading a 4GB AI model onto users’ devices without requesting permission, raising fresh concerns around privacy, storage usage, and user consent. According to thatprivacyguy, the Gemini Nano weights file is automatically stored inside the “OptGuideOnDeviceModel” directory without any prompt, notification, or approval from users.

The report claims that millions of Chrome users may now unknowingly have the AI model stored on their systems, consuming valuable storage space in the background. Critics argue that the large-scale deployment also carries a significant environmental impact due to the energy required for transferring and storing such massive files globally.

One of the major concerns highlighted is the difficulty users face when trying to remove the file. The report states that manually deleting the “weights.bin” file does not permanently solve the issue, as Chrome allegedly downloads the file again the next time the browser launches.

For Windows users, permanently stopping the download reportedly requires editing the system registry. Mac users, meanwhile, need to disable specific Chrome flags manually through browser settings.

The article further points toward possible legal concerns under European privacy regulations. Under Article 5(3) of the EU’s ePrivacy Directive, companies are required to obtain “prior, freely-given, specific, informed, and unambiguous consent” before storing information on a user’s device. Privacy experts cited in the report believe Chrome’s silent installation process may conflict with these requirements.

Environmental concerns were also raised in the report, which estimates that transferring AI models of this size across Chrome’s enormous user base could consume electricity on a massive scale. The article argues that repeated downloads caused by users attempting to delete the files could further increase the overall carbon footprint.

Users who wish to disable the feature reportedly need to take technical steps. Windows users can navigate to “HKEY_LOCAL_MACHINE\SOFTWARE\Policies\Google\Chrome,” create a DWORD named “GenAILocalFoundationalModelSettings,” set its value to 1, and restart Chrome. Mac users can disable the feature through chrome://flags by turning off “Enables Optimization Guide On Device.”

The incident has sparked broader debates around digital consent, AI deployment practices, and whether tech companies should be allowed to push large software components onto personal devices without explicit user approval.

Ubuntu Services Remain Disrupted After DDoS Attack Targets Canonical Infrastructure



 

website

Cyber Attacks DDOS Attack ransomware website

Ubuntu Services Remain Disrupted After DDoS Attack Targets Canonical Infrastructure

Several Ubuntu users reported problems installing updates and downloading packages after parts of Canonical’s infrastructure were disrupted during a Distributed Denial of Service (DDoS) attack. Canonical, the company behind the Ubuntu Linux distribution, confirmed that its online systems had been targeted.

In a statement released during the outage, Canonical said its web infrastructure was facing what it described as a sustained cross-border cyberattack and that teams were working to restore affected services. The company added that further updates would be shared through official channels once more information became available.

Discussions across Ubuntu community forums suggested that multiple services were affected during the incident, including Ubuntu’s security API and several Canonical-operated websites. Users also stated that software installations and system updates were temporarily unavailable or failing to complete properly.

Responsibility for the attack was later claimed by a group calling itself “The Islamic Cyber Resistance in Iraq 313 Team.” In Telegram posts attributed to the group, the attackers allegedly said they used a DDoS-for-hire platform known as “Beamed” to carry out the operation.

Beamed is described as a “booter” or “stresser” service, which are platforms that allow customers to pay for DDoS attacks. These services are often advertised as tools for testing website traffic capacity, although security researchers have repeatedly linked them to disruptive cyber operations. According to claims associated with the platform, Beamed is capable of generating attacks reaching 3.5 terabits per second, enough traffic to overwhelm major online infrastructure.

A DDoS attack works by flooding a server or network with enormous volumes of internet traffic from large numbers of connected devices at the same time. Once systems become overloaded, legitimate users may no longer be able to access websites, applications, or online services. Unlike ransomware campaigns or data breaches, the primary goal of most DDoS attacks is to interrupt availability rather than steal information directly.

To create these attack networks, threat actors typically compromise internet-connected devices using malware. Weak passwords, exposed systems, outdated software, and poorly secured smart devices are commonly targeted. Once infected, the devices become part of a botnet that can be remotely controlled through centralized management panels.

Access to these botnets is frequently sold through underground marketplaces and subscription-based services. Depending on the size and duration of the attack, prices can range from as little as $10 for lower-powered services to hundreds of dollars per month for larger and more persistent attacks.

The disruption drew attention within the open-source community because Ubuntu infrastructure is widely used across enterprise servers, development environments, cloud systems, and research institutions worldwide. Problems affecting package repositories or security update services can delay software deployments and patch management for organizations that rely on Ubuntu systems daily.

The incident also reflects how accessible DDoS-for-hire services have become over the past few years. Platforms offering attack infrastructure continue to reduce the technical barrier required to launch disruptive cyberattacks, allowing even low-skilled actors to rent large-scale attack capabilities for relatively small amounts of money.

Critical OpenClaw Flaws Allow Persistent Access and Credential Abuse



 

WebSocket Exploits

AI Agent Security AI Runtime Vulnerabilities Browser Session Hijacking Credential Theft Cyber Security OpenClaw Vulnerabilities Privilege Escalation Prompt Injection Attacks WebSocket Exploits

Critical OpenClaw Flaws Allow Persistent Access and Credential Abuse

OpenClaw, a self-hosted AI agent runtime which has gained rapid adoption by enterprises, introduces a new type of security exposure for enterprises as dynamically executed content, external skill integrations, and cloud-based authentication mechanisms are convergent without adequate defensive control mechanisms.

The OpenClaw platform is unlike conventional applications that are constructed using fixed execution logic, as it is capable of accepting untrusted inputs, retrieving and executing third-party code modules, and interacting with connected environments with assigned credentials, effectively extending the trust boundary far beyond the application layer itself. These architectural flexibility and the recently disclosed ClawJacked exploitation technique expose critical weaknesses in authentication handling and token protection within browser-based cloud development environments, according to security researchers.

It has been demonstrated that malicious web content can exploit active developer sessions to extract sensitive access tokens, thereby granting attackers unauthorized access to source repositories, cloud infrastructures, and privileged enterprise resources. Increasingly, organizations are integrating cloud-native development platforms into their engineering workflows. This disclosure highlights concerns regarding privilege scoping, identity isolation, and other security aspects associated with autonomous AI-powered runtime environments.

A coordinated vulnerability chain, collectively known as the "Claw Chain," was identified by Cyera researchers in response to these concerns, demonstrating how multiple vulnerabilities within OpenClaw can be combined to compromise a system, gain unauthorized access to data, and escalate privileges across affected systems.

In particular, two vulnerabilities have been assigned CVE-2026-44113 and CVE-2026-2026-44112, which contain time-of-check/time-of-use (TOCTOU) race conditions within the OpenShell managed sandbox backend, which could allow attackers to circumvent sandbox enforcement and interact with files outside of the mounted root.

In contrast to the first issue, which permits arbitrary write operations which can lead to configuration changes, backdoor installations, and long-term control over compromised hosts, the second issue provides a pathway for unauthorized disclosure of system artifacts, credentials, and sensitive internal data through unauthorized file disclosure.

Researchers also disclosed CVE-2026-44115, a vulnerability resulting from an incomplete denylist implementation that allows adversaries to conceal shell expansion tokens in heredoc payloads and execute commands that bypass runtime restrictions.

A fourth vulnerability known as CVE-2026-44118 introduces an improper access control condition in which non-owner loopback clients can impersonate privileged users to manipulate gateway configurations, alter scheduled cron operations, and gain greater control of execution environments through unauthorized use of privileged accounts. These flaws collectively demonstrate the possibility of insufficient isolation, weak privilege boundaries, and inadequate runtime validation mechanisms within modern AI agent infrastructures resulting in a full compromise chain which can sustain stealthy and persistent access despite seemingly isolated weaknesses.

OpenClaw's rapid adoption and permissive architecture have contributed to its rapid transformation from a niche automation framework into a widely deployed AI-driven orchestration environment, further amplifying its security implications.

In late 2025, Austrian engineer Peter Steinberger released a public version of the project that gained wide traction because of its unique capability to provide custom automation capabilities outside of tightly controlled commercial ecosystems. The OpenClaw assistant does not rely on vendor-defined integrations, but rather allows users to develop, modify, and distribute executable "skills."

The result is a large repository containing thousands of automation scenarios developed by the community without centrally managing, categorizing, or validating their security. Due to its “self-hackability” design, where configurations, memory stores, and executable logic are maintained using local Markdown-based structures that can be modified by the user, it has attracted both developer interest and growing scrutiny from security researchers concerned about the absence of hardened trust boundaries.

It was discovered that hundreds of OpenClaw administrative interfaces were accessible over the internet and did not require authentication. These concerns escalated. Investigations revealed that improperly configured reverse proxies could forward external traffic through localhost-trusted channels, causing the platform to mistakenly treat remote requests as privileged local connections.

Security researcher Jamieson O'Reilly demonstrated the severity of the issue by gaining access to sensitive assets such as credentials for Anthropic APIs, Telegram bot tokens, Slack environments, and archived conversations. Further research revealed that prompt injection attacks could be used to manipulate the agent to perform unintended behavior by embedding malicious instructions in emails, files, or web content processed by the underlying large language model.

One such scenario was demonstrated by Matvey Kukuy's delivery of crafted email payloads which coerced the bot to provide private cryptographic keys from the host environment upon receiving instructions to review inbox contents. Several independent experiments have demonstrated the system discloses confidential email data, exposes the contents of home directories via automated shell commands, and searches local storage automatically after receiving psychologically manipulative prompts.

In aggregate, these incidents illustrate an industry concern that autonomous AI agents operating with wide filesystem visibility, persistent memory, and delegated execution privileges may be highly susceptible to indirect command manipulation when deployed in a manner that does not adhere to strict authentication controls, runtime isolation, and contextual validation controls.

Despite the fact that there is no publicly verified link to any known advanced persistent threat group linking the exploitation of the OpenClaw vulnerabilities, security analysts note that the operational characteristics of the attack are in line with tradecraft commonly utilized in credential theft, browser hijacking, and adversary-in-the-middle intrusion campaigns.

MITRE ATT&CK framework techniques, including T1185 related to browser session hijacking as well as T1557 related to man-in-the-middle attacks, have been identified as parallel techniques, and both of these techniques are frequently used in targeted attacks against enterprise authentication systems and cloud-based environments. There has been a growing concern that financially motivated threat actors and state-aligned operators may incorporate the technique into broader intrusion toolsets due to the availability of publicly available proof-of-concept exploit methods and the relatively low complexity required to weaponize these flaws.

It was discovered that all versions of OpenClaw and Clawdbot before version 2026.2.2, including all builds up to version 2026.2.1, have been vulnerable to the vulnerability. Researchers stated that in the updated version, unauthorized WebSocket interactions are restricted and authentication checks are enforced on the exposed /cdp interface, which previously permitted unsafe assumptions regarding local trust.

During the deployment of immediate patches, security teams are advised to monitor for suspicious localhost WebSocket activity, unauthorized browser extension behaviors, and attempts to communicate outbound via ws://127.0.0.1:17892/cdp or infrastructure controlled by known attackers.

When rapid patching is an operational challenge, experts recommend that the OpenClaw browser extension be temporarily disabled, that host-level firewall restrictions be enforced around local WebSocket services, and that browser session telemetry and endpoint indicators of compromise be continuously reviewed to determine if there has been an unauthorized persistence of credentials or credential interception.

OpenClaw's vulnerability chain is a reflection of an overall security reckoning taking place in the rapidly expanding AI agent ecosystem, in which convenience-driven automation is outpacing the maturation of defensive safeguards designed to contain it in a rapidly expanding ecosystem. There is an increasing tendency for autonomous assistants to gain access to developer environments, authentication tokens, local storage, messaging platforms, and cloud infrastructure, so that the traditional boundaries between trusted execution and untrusted input are being eroded.

Platforms with the ability to self-modify, delegate command execution, and persist contextual memory present significant security risks that are fundamentally different from conventional software, particularly when deployed with excessive privileges and inadequate isolation during runtime.

Despite the fact that OpenClaw's vulnerabilities may be mitigated by patching, access restrictions, and stronger authentication enforcement, the incident emphasizes the larger industry concern that artificial intelligence-driven operational tools may become a high value target for both cybercriminals and advanced intrusion groups in the very near future.

These findings serve as a reminder that, as organizations adopt autonomous AI systems, security architecture, privilege segmentation, and continuous monitoring must no longer be overlooked.

Search This Blog

Sections

Popular Posts

Blog Archive

Labels

About Me

Footer About

Labels

Showing result(s) for

Popular Posts

Pages

Menu Item

About quantum computing and threats

Why is quantum computing so fast?

Why are blockchains weak against quantum computing?

Why is crypto migration difficult?

Is the digital industry prepared?

2026 and cybersecurity

Is the convergence good or bad?

1. System and skills problem

2. Bug management shift to active exposure reduction

3. Advanced deepfake protection is the need of the hour

4. Post-quantum protection

Private company reported the incident immediately

What is RBI’s zero liability rule?

Bank will return stolen amount with interest