2026-03-09 AI创业新闻

OpenAI Codex Security Scanned 1.2 Million Commits and Found 10,561 High-Severity Issues

OpenAI on Friday began rolling out Codex Security , an artificial intelligence (AI)-powered security agent that’s designed to find, validate, and propose fixes for vulnerabilities. The feature is available in a research preview to ChatGPT Pro, Enterprise, Business, and Edu customers via the Codex web with free usage for the next month. “It builds deep context about your project to identify complex vulnerabilities that other agentic tools miss, surfacing higher-confidence findings with fixes that meaningfully improve the security of your system while sparing you from the noise of insignificant bugs,” the company said . Codex Security represents an evolution of Aardvark⁠ , which OpenAI unveiled in private beta in October 2025 as a way for developers and security teams to detect and fix security vulnerabilities at scale.

Over the last 30 days, Codex Security has scanned more than 1.2 million commits across external repositories over the course of the beta, identifying 792 critical findings and 10,561 high-severity findings. These include vulnerabilities in various open-source projects like OpenSSH⁠, GnuTLS⁠, GOGS⁠, Thorium⁠, libssh, PHP, and Chromium, among others. Some of them have been listed below - GnuPG - CVE-2026-24881, CVE-2026-24882 GnuTLS - CVE-2025-32988, CVE-2025-32989 GOGS - CVE-2025-64175, CVE-2026-25242 Thorium

CVE-2025-35430, CVE-2025-35431, CVE-2025-35432, CVE-2025-35433, CVE-2025-35434, CVE-2025-35435, CVE-2025-35436 According to the AI company, the latest iteration of the application security agent leverages the reasoning capabilities of its frontier models and combines them with automated validation to minimize the risk of false positives and deliver actionable fixes. OpenAI’s scans on the same repositories over time have demonstrated increasing precision and declining false positive rates, with the latter falling by more than 50% across all repositories.

In a statement shared with The Hacker News, OpenAI said Codex Security is designed to improve signal-to-noise by grounding vulnerability discovery in system context and validating findings before surfacing them to users. Specifically, the agent works in three steps: it analyzes a repository to get a handle on the project’s security-relevant structure of the system and generates an editable threat model that captures what it does and where it’s most exposed. Once the system context is built, Codex Security uses it as a foundation to identify vulnerabilities and classifies findings based on their real-world impact. The flagged issues are pressure-tested in a sandboxed environment to validate them.

“When Codex Security is configured with an environment tailored to your project, it can validate potential issues directly in the context of the running system,” OpenAI said. “That deeper validation can reduce false positives even further and enable the creation of working proofs-of-concept, giving security teams stronger evidence and a clearer path to remediation.” The final stage involves the agent proposing fixes that best align with the system behavior so as to reduce regressions and make them easier to review and deploy. News of Codex Security comes weeks after Anthropic launched Claude Code Security to help users scan a software codebase for vulnerabilities and suggest patches. Found this article interesting?

Anthropic Finds 22 Firefox Vulnerabilities Using Claude Opus 4.6 AI Model

Anthropic on Friday said it discovered 22 new security vulnerabilities in the Firefox web browser as part of a security partnership with Mozilla. Of these, 14 have been classified as high, seven have been classified as moderate, and one has been rated low in severity. The issues were addressed in Firefox 148 , released late last month. The vulnerabilities were identified over a two-week period in January 2026.

The artificial intelligence (AI) company said the number of high-severity bugs identified by its Claude Opus 4.6 large language model (LLM) represents “almost a fifth” of all high-severity vulnerabilities that were patched in Firefox in 2025. Anthropic said the LLM detected a use-after-free bug in the browser’s JavaScript after “just” 20 minutes of exploration, which was then validated by a human researcher in a virtualized environment to rule out the possibility of a false positive. “By the end of this effort, we had scanned nearly 6,000 C++ files and submitted a total of 112 unique reports, including the high- and moderate-severity vulnerabilities mentioned above,” the company said. “Most issues have been fixed in Firefox 148, with the remainder to be fixed in upcoming releases.” The AI upstart said it also fed its Claude model access to the entire list of vulnerabilities submitted to Mozilla and tasked the AI tool with developing a practical exploit for them.

Despite carrying out the test several hundred times and spending about $4,000 in API credits, the company said Claude Opus 4.6 was able to turn the security defect into an exploit only in two cases. This behavior, the company added, signaled two important aspects: the cost of identifying vulnerabilities is cheaper than creating an exploit for them, and the model is better at finding issues than at exploiting them. “However, the fact that Claude could succeed at automatically developing a crude browser exploit, even if only in a few cases, is concerning,” Anthropic emphasized, adding the exploits only worked within the confines of its testing environment, which has had some security features like sandboxing intentionally stripped off. A crucial component incorporated into the process is a task verifier to determine if the exploit actually works, giving the tool real-time feedback as it explores the codebase in question and allowing it to iterate its results until a successful exploit is devised.

One such exploit Claude wrote was for CVE-2026-2796 (CVSS score: 9.8), which has been described as a just-in-time (JIT) miscompilation in the JavaScript WebAssembly component. The disclosure comes weeks after the company released Claude Code Security in a limited research preview as a way to fix vulnerabilities using an AI agent. “We can’t guarantee that all agent-generated patches that pass these tests are good enough to merge immediately,” Anthropic said. “But task verifiers give us increased confidence that the produced patch will fix the specific vulnerability while preserving program functionality—and therefore achieve what’s considered to be the minimum requirement for a plausible patch.” Mozilla, in a coordinated announcement, said the AI-assisted approach has discovered 90 other bugs, most of which have been fixed.

These consisted of assertion failures that overlapped with issues traditionally found through fuzzing and distinct classes of logic errors that the fuzzers failed to catch. “The scale of findings reflects the power of combining rigorous engineering with new analysis tools for continuous improvement,” the browser maker said . “We view this as clear evidence that large-scale, AI-assisted analysis is a powerful new addition to security engineers’ toolbox.” Found this article interesting? Follow us on Google News , Twitter and LinkedIn to read more exclusive content we post.

Transparent Tribe Uses AI to Mass-Produce Malware Implants in Campaign Targeting India

The Pakistan-aligned threat actor known as Transparent Tribe has become the latest hacking group to embrace artificial intelligence (AI)-powered coding tools to strike targets with various implants. The activity is designed to produce a “high-volume, mediocre mass of implants” that are developed using lesser-known programming languages like Nim, Zig, and Crystal and rely on trusted services like Slack, Discord, Supabase, and Google Sheets to fly under the radar, according to new findings from Bitdefender. “Rather than a breakthrough in technical sophistication, we are seeing a transition toward AI-assisted malware industrialization that allows the actor to flood target environments with disposable, polyglot binaries,” security researchers Radu Tudorica, Adrian Schipor, Victor Vrabie, Marius Baciu, and Martin Zugec said in a technical breakdown of the campaign. The transition towards vibe-coded malware, aka vibeware , as a means to complicate detection has been characterized by the Romanian cybersecurity vendor as Distributed Denial of Detection (DDoD).

In this approach, the idea is not to sidestep detection efforts through technical sophistication, but rather to flood target environments with disposable binaries, each using a different language and communication protocol. Helping threat actors in this aspect are large language models (LLMs), which lower the barrier to cybercrime and collapse the expertise gap by enabling them to generate functional code in unfamiliar languages, either from scratch or by porting the core business logic from more common ones. The latest set of attacks has been found to target the Indian government and its embassies in multiple foreign countries, with APT36 using LinkedIn to identify high-value targets. The attacks have also singled out the Afghan government and several private businesses, albeit to a lesser extent.

The infection chains likely begin with phishing emails bearing Windows shortcuts (LNKs) bundled within ZIP archives or ISO images. Alternatively, PDF lures featuring a prominent “Download Document” button are used to redirect users to an attacker-controlled website that triggers the download of the same ZIP archives. Regardless of the method used, the LNK file is used to execute PowerShell scripts in memory, which then download and run the main backdoor and facilitate post-compromise actions. These include the deployment of known adversary simulation tools like Cobalt Strike and Havoc, indicating a hybrid approach to ensure resilience.

Some of the other tools observed as part of the attacks are listed below - Warcode , a custom shellcode loader written in Crystal that’s used to reflectively load a Havoc agent directly into memory. NimShellcodeLoader , an experimental counterpart to Warcode that’s used to deploy a Cobalt Strike beacon embedded into it. CreepDropper , a .NET malware that’s used to deliver and install additional payloads, including SHEETCREEP, a Go-based infostealer that uses Microsoft Graph API for C2, and MAILCREEP, a C#-based backdoor utilizing Google Sheets for C2. Both malware families were detailed by Zscaler ThreatLabz in January 2026.

SupaServ , a Rust-based backdoor that establishes a primary communication channel via the Supabase platform, with Firebase acting as a fallback. It contains Unicode emojis, suggesting that it was likely developed using AI. LuminousStealer , a likely vibe-coded, Rust-based infostealer that uses Firebase and Google Drive to exfiltrate files matching certain extensions (.txt, .docx, .pdf, .png, .jpg, .xlsx, .pptx, .zip, .rar, .doc, and .xls). CrystalShell , a backdoor written in Crystal that’s capable of targeting Windows, Linux, and macOS systems, and uses hard-coded Discord channel IDs for C2.

It supports the ability to run commands and gather host information. One variant of the malware has been found to use Slack for C2. ZigShell , a counterpart to CrystalShell that’s written in Zig and uses Slack as its primary C2 infrastructure. It also supports added functionality to upload and download files.

CrystalFile , a simple command interpreter written in Crystal that continuously monitors the “C:\Users\Public\AccountPictures\input.txt” and executes the contents using “cmd.exe.” LuminousCookies , a Rust-based specialized injector to exfiltrate cookies, passwords, and payment information from Chromium-based browsers by circumventing app-bound encryption . BackupSpy , a Rust-based utility designed to monitor the local file system and external media for high-value data. ZigLoader , a specialized loader written in Zig that decrypts and executes arbitrary shellcode in memory. Gate Sentinel Beacon , a customized version of the open-source GateSentinel C2 framework project.

“The transition of APT36 toward vibeware represents a technical regression,” Bitdefender said. “While AI-assisted development increases sample volume, the resulting tools are often unstable and riddled with logical errors. The actor’s strategy incorrectly targets signature-based detection, which has long been superseded by modern endpoint security.” Bitdefender haș warned that the threat posed by AI-assisted malware is the industrialization of the attacks, allowing threat actors to scale their activities quickly and with less effort. “We are seeing a convergence of two trends that have been developing for some time: the adoption of exotic, niche programming languages, and the abuse of trusted services to hide in legitimate network traffic,” the researchers said.

“This combination allows even mediocre code to achieve high operational success by simply overwhelming standard defensive telemetry.” Found this article interesting? Follow us on Google News , Twitter and LinkedIn to read more exclusive content we post.

Multi-Stage VOID#GEIST Malware Delivering XWorm, AsyncRAT, and Xeno RAT

Cybersecurity researchers have disclosed details of a multi-stage malware campaign that uses batch scripts as a pathway to deliver various encrypted remote access trojan (RATs) payloads that correspond to XWorm , AsyncRAT , and Xeno RAT . The stealthy attack chain has been codenamed VOID#GEIST by Securonix Threat Research. At a high level, the obfuscated batch script is used to deploy a second batch script, stage a legitimate embedded Python runtime, and decrypt encrypted shellcode blobs, which are executed directly in memory by injecting them into separate instances of “explorer.exe” using a technique called Early Bird Asynchronous Procedure Call (APC) injection . “Modern malware campaigns increasingly shift from standalone executables toward complex, script-based delivery frameworks that closely mimic legitimate user activity,” researchers Akshay Gaikwad, Shikha Sangwan, and Aaron Beardslee said in a technical report shared with The Hacker News.

“Rather than deploying traditional PE binaries, attackers leverage modular pipelines comprising batch scripts for orchestration, PowerShell for stealthy staging, legitimate embedded runtimes for portability, and raw shellcode executed directly in memory for persistence and control.” This fileless execution mechanism minimizes disk-based detection opportunities, thereby allowing the threat actors to operate within compromised systems without triggering security alerts. What’s more, the approach offers an extra advantage in that these individual stages appear harmless in isolation and resemble regular administrative activity. The starting point of the attack is a batch script that’s fetched from a TryCloudflare domain and distributed via phishing emails. Once launched, it deliberately avoids taking steps to escalate privileges and leverages the permission rights of the currently logged-in user to establish an initial foothold, while blending into seemingly innocuous administrative operations.

The initial stage serves as a launchpad to display a decoy PDF by launching Google Chrome in full-screen. The displayed financial document or invoice serves as a visual distraction to conceal what’s happening behind the scenes. This includes launching a PowerShell command to re-execute the original batch script, such as using the -WindowStyle Hidden parameter, to avoid displaying a console window. To ensure persistence across system reboots, an auxiliary batch script is placed in the Windows user’s Startup directory so that it’s automatically executed every time the victim logs in to the system.

The absence of more intrusive persistence methods is intentional, as it reduces the forensic footprint. “Technically, this persistence method operates entirely within the current user’s privilege context. It does not modify system-wide registry keys, create scheduled tasks, or install services,” the researchers said. “Instead, it relies on standard user-level startup behavior, which requires no elevation and generates minimal security friction.

This design choice reduces the likelihood of triggering privilege escalation prompts or registry-monitoring alerts.” The next phase begins with the malware reaching out to a TryCloudflare domain to fetch additional payloads in the form of ZIP archives that contain multiple files - runn.py , a Python-based loader script responsible for decrypting and injecting encrypted shellcode payload modules into memory new.bin , an encrypted shellcode payload corresponding to XWorm xn.bin , an encrypted shellcode payload corresponding to Xeno RAT pul.bin , an encrypted shellcode payload corresponding to AsyncRAT a.json, n.json, and p.json , key files containing the decryption keys required by the Python loader to dynamically decrypt the shellcode at runtime Once the files are extracted, the attack sequence deploys a legitimate embedded Python runtime directly from python[.]org. This step offers several advantages. For starters, it eliminates any dependency on the system. As a result, the malware can continue to operate even if the infected endpoint does have Python installed.

“From the attacker’s perspective, the objectives of this stage are portability, reliability, and stealth,” Securonix said. “By embedding a legitimate interpreter into the staging directory, the malware transforms itself into a fully self-contained execution environment capable of decrypting and injecting payload modules without relying on external system components.” The main goal of the attack is to leverage the Python runtime to launch “runn.py,” which then decrypts and runs the XWorm payload using Early Bird APC injection. The malware also makes use of a legitimate Microsoft binary, “AppInstallerPythonRedirector.exe,” to invoke Python and launch Xeno RAT. In the last stage, the Python loader uses the same injection mechanism to launch AsyncRAT.

The infection chain culminates with the malware transmitting a minimal HTTP beacon back to attacker-controlled C2 infrastructure hosted on TryCloudflare to confirm the digital break-in. It’s currently not known who the targets of the attack were, and if there have been any successful compromises. “This repeated injection pattern reinforces the modular architecture of the framework. Instead of delivering a single monolithic payload, the attacker deploys components incrementally, improving flexibility and resilience,” Securonix said.

“From a detection standpoint, repeated process injection into explorer.exe within short time windows is a strong behavioral indicator that correlates across stages of the attack.” Found this article interesting? Follow us on Google News , Twitter and LinkedIn to read more exclusive content we post.

The MSP Guide to Using AI-Powered Risk Management to Scale Cybersecurity

Scaling cybersecurity services as an MSP or MSSP requires technical expertise and a business model that delivers measurable value at scale. Risk-based cybersecurity is the foundation of that model. When done right, it builds client trust, increases upsell opportunities, and drives recurring revenue. But to deliver this consistently and efficiently, you need the right technology and processes.

We created The MSP Growth Guide: How MSPs Use AI-Powered Risk Management to Scale Their Cybersecurity Business to help providers transition to scalable, risk-first cybersecurity. Inside, you’ll find practical insights into the top challenges MSPs face, expert guidance on overcoming them, and a framework for selecting and implementing AI-powered risk management to unlock scalable, recurring revenue. Why Risk Management Is the Key to Scaling Cybersecurity Services Most MSPs offer critical cybersecurity services, from compliance support to endpoint protection, but these are often isolated engagements that limit long-term value and recurring revenue. A risk-based approach changes that.

By assessing the full threat landscape and prioritizing risks by business impact, MSPs can shift from tactical fixes to continuous, proactive service. By adopting a risk‑first approach, you can: Anticipate and neutralize threats before they cause harm Continuously adapt security measures to an evolving threat landscape Protect assets, operations, and reputation even when compliance does not require specific actions Risk management also helps MSPs meet the expectations of modern cybersecurity frameworks, many of which require formal, ongoing risk assessments. By embedding risk management into your service offerings, you open the door to more lucrative contracts and compliance‑led upsells. Six Challenges AI-Powered Risk Management Solves for MSPs Offering risk management services delivers clear value, yet even experienced MSPs encounter obstacles that hinder service delivery, reduce scalability, and make it more difficult to showcase their impact to clients.

Here are the six most common barriers to growth MSPs face:
Manual assessments: Time-consuming, error-prone, and hard to scale No remediation roadmap; Findings without clear action plans frustrate clients Compliance complexity; Aligning with multiple frameworks manually is time-consuming and inconsistent Lack of business context; Reports are too technical for decision-makers Talent shortages; Skilled risk experts are hard to find and retain Unmanaged third-party risk; Most platforms ignore vendor risk To turn a risk-based cybersecurity strategy into a scalable, profitable service model, MSPs need the right technology. That’s where AI-powered risk management platforms come in. These platforms streamline every step, from assessment to remediation and reporting, while embedding CISO-level expertise into your service delivery. Selecting the Right AI‑Powered Risk Management Platform — What to Demand and Why What to Expect from a Modern Risk Management Platform The right AI‑powered risk management platform assesses threats while accelerating the delivery of results that drive business growth.

Service providers should expect: Faster onboarding and service delivery with automated, user-friendly risk assessments Improved compliance management through built-in framework alignment, automatic mapping, and continuous monitoring Higher client satisfaction and trust with clear, business-focused risk reporting Measurable ROI by reducing manual workloads, increasing efficiency, and enabling more profitable service delivery at scale Greater upsell opportunities by identifying additional services clients need based on their unique risk profile How to Choose the Right AI-Powered Risk Management Platform Choosing the right risk management solution is key to scaling cybersecurity services. The right platform should improve operational efficiency, help prioritize action, and communicate risk in ways that resonate with business stakeholders. Key capabilities to look for in a modern risk management platform include: Automated risk assessments: Deliver results in days instead of months, reduce human error, and ensure consistent, repeatable outcomes Dynamic risk register with heatmaps: Instantly visualize and prioritize risks based on severity and likelihood to focus efforts where they matter most Actionable remediation plans: Turn findings into clear, prioritized tasks aligned with business goals and compliance requirements Customizable risk tolerances: Tailor risk scoring and recommendations to each client’s specific objectives and risk appetite For the full list of capabilities to look for, download The MSP Growth Guide: How MSPs Use AI-Powered Risk Management to Scale Their Cybersecurity Business. When these capabilities are built into your service model, you unlock the ability to scale operations, increase margins, and grow revenue without expanding headcount.

Turning AI-Powered Risk Management Into a Strategic Advantage AI-powered risk management helps MSPs and MSSPs scale services, improve efficiency, and deliver continuous value to clients. It streamlines assessments, prioritizes risks based on business impact, and supports consistent, high-quality service delivery. The MSP Growth Guide: How MSPs Use AI-Powered Risk Management to Scale Their Cybersecurity Business shows how to integrate AI-driven risk management into your offering to drive long-term growth. Download the guide to learn how to scale smarter, strengthen client relationships, and use risk management as a competitive advantage.

Found this article interesting? This article is a contributed piece from one of our valued partners. Follow us on Google News , Twitter and LinkedIn to read more exclusive content we post.

Accelerate your AI Initiatives

Iran-Linked MuddyWater Hackers Target U.S. Networks With New Dindoor Backdoor

New research from Broadcom’s Symantec and Carbon Black Threat Hunter Team has discovered evidence of an Iranian hacking group embedding itself in several U.S. companies’ networks, including banks, airports, non-profit, and the Israeli arm of a software company. The activity has been attributed to a state-sponsored hacking group called MuddyWater (aka Seedworm). It’s affiliated with the Iranian Ministry of Intelligence and Security (MOIS).

The campaign is assessed to have begun in early February, with recent activity detected following U.S. and Israeli military strikes on Iran . “The software company is a supplier to the defense and aerospace industries, among others, and has a presence in Israel, with the company’s Israel operation seeming to be the target in this activity,” the security vendor said in a report shared with The Hacker News. The attacks targeting the software company, as well as a U.S.

bank and a Canadian non-profit, have been found to pave the way for a previously unknown backdoor dubbed Dindoor, which leverages the Deno JavaScript runtime for execution. Broadcom said it also identified an attempt to exfiltrate data from the software company using the Rclone utility to a Wasabi cloud storage bucket. However, it’s currently not known if the effort paid off. Also found in the networks of a U.S.

airport and a non-profit was a separate Python backdoor called Fakeset, which was downloaded from servers belonging to Backblaze, an American cloud storage and data backup company. The digital certificate used to sign Fakeset has also been used to sign Stagecomp and Darkcomp malware, both previously linked to MuddyWater. “While this malware wasn’t seen on the targeted networks, the use of the same certificates suggests the same actor – namely Seedworm – was behind the activity on the networks of the U.S. companies,” Symantec and Carbon Black said.

“Iranian threat actors have become increasingly proficient in recent years. Not only has their tooling and malware improved, but they’ve also demonstrated strong social engineering capabilities, including spear-phishing campaigns and ‘honeytrap’ operations used to build relationships with targets of interest to gain access to accounts or sensitive information.” The findings come against the backdrop of an escalating military conflict in Iran, triggering a barrage of cyber attacks in the digital sphere. Recent research from Check Point has uncovered the pro-Palestinian hacktivist group known as Handala Hack (aka Void Manticore) routing its operations through Starlink IP ranges to probe externally facing applications for misconfigurations and weak credentials. In recent months, multiple Iran-nexus adversaries , such as Agrius (aka Agonizing Serpens, Marshtreader, and Pink Sandstorm), have also observed scanning for vulnerable Hikvision cameras and video intercom solutions using known security flaws such as CVE-2017-7921 and CVE-2023-6895 .

The targeting, per Check Point, has intensified in the wake of the current Middle East conflict. The exploitation attempts against IP cameras have witnessed a surge in Israel and Gulf countries, including the U.A.E., Qatar, Bahrain, and Kuwait, along with Lebanon and Cyprus. The activity has singled out cameras from Dahua and Hikvision, weaponizing the two aforementioned vulnerabilities, as well as CVE-2021-36260 , CVE-2025-34067 , and CVE-2021-33044 . “Taken together, these findings are consistent with the assessment that Iran, as part of its doctrine, leverages camera compromise for operational support and ongoing battle damage assessment (BDA) for missile operations, potentially in some cases prior to missile launches,” the company said .

“As a result, tracking camera-targeting activity from specific, attributed infrastructures may serve as an early indicator of potential follow-on kinetic activity.” The U.S. and Israel’s war with Iran has also prompted an advisory from the Canadian Centre for Cyber Security (CCCS), which cautioned that Iran will likely use its cyber apparatus to stage retaliatory attacks against critical infrastructure and information operations to further the regime’s interests. Some other key developments that have unfolded in recent days are listed below - Israeli intelligence agencies hacked into Tehran’s extensive traffic camera network for years to monitor the movements of bodyguards of Ayatollah Ali Khamenei and other top Iranian officials in the lead up to the assassination of the supreme leader last week, the Financial Times reported . Iran’s Islamic Revolutionary Guard Corps (IRGC) targeted Amazon’s data center in Bahrain for the company’s support of the “enemy’s military and intelligence activities,” state media Fars News Agency said on Telegram.

Active wiper campaigns are said to be underway against Israeli energy, financial, government, and utilities sectors. “Iran’s wiper arsenal includes 15+ families (ZeroCleare, Meteor, Dustman, DEADWOOD, Apostle, BFG Agonizer, MultiLayer, PartialWasher, and others),” Anomali said . Iranian state-sponsored APT groups like MuddyWater, Charming Kitten, OilRig, Elfin, and Fox Kitten “demonstrated clear signs of activation and rapid retooling, positioning themselves for retaliatory operations amid the escalating conflict,” LevelBlue said , adding “cyber represents one of Iran’s most accessible asymmetric tools for retaliation against Gulf states that condemned its attacks and support U.S. operations.” According to Flashpoint, a massive #OpIsrael cyber campaign involving pro-Russian and pro-Iranian actors has targeted Israeli industrial control systems (ICS) and government portals across Kuwait, Jordan, and Bahrain.

The campaign is driven by NoName057(16), Handala Hack, Fatemiyoun Electronic Team, and Cyber Islamic Resistance (aka 313 Team). Between 28 February 2026 and 2 March 2026, pro-Russia hacktivist group Z-Pentest claimed responsibility for compromising several U.S.-based entities, including ICS and SCADA systems and multiple CCTV networks. “The timing of these unverified claims, coinciding with Operation Epic Fury, suggests Z-Pentest likely began prioritizing U.S. entities as targets,” Adam Meyers, head of Counter Adversary Operations at CrowdStrike, told The Hacker News.

“Iran’s offensive cyber capability has matured into a durable instrument of state power used to support intelligence collection, regional influence, and strategic signaling during periods of geopolitical tension,” UltraViolet Cyber said . “A defining feature of Iran’s current cyber doctrine is its emphasis on identity and cloud control planes as the primary attack surface.” “Rather than prioritizing zero-day exploitation or highly novel malware at scale, Iranian operators tend to focus on repeatable access techniques such as credential theft, password spraying, and social engineering, followed by persistence through widely deployed enterprise services.” Organizations are advised to bolster their cybersecurity posture, strengthen monitoring capabilities, limit exposure to the internet, disable remote access to operational technology (OT) systems, enforce phishing-resistant multi-factor authentication (MFA), implement network segmentation, take offline backups, and ensure that all internet-facing applications, VPN gateways, and edge devices are up-to-date “Western organizations should continue to remain on high-alert for potential cyber response as the conflict continues and activity may move beyond hacktivism and into destructive operations,” Meyers said. Found this article interesting? Follow us on Google News , Twitter and LinkedIn to read more exclusive content we post.

China-Linked Hackers Use TernDoor, PeerTime, BruteEntry in South American Telecom Attacks

A China-linked advanced persistent threat (APT) actor has been targeting critical telecommunications infrastructure in South America since 2024, targeting Windows and Linux systems and edge devices with three different implants. The activity is being tracked by Cisco Talos under the moniker UAT-9244 , describing it as closely associated with another cluster known as FamousSparrow . It’s worth noting that FamousSparrow is assessed to share tactical overlaps with Salt Typhoon , a China-nexus espionage group known for its targeting of telecommunication service providers. Despite the similar targeting footprint between UAT-9244 and Salt Typhoon, there is no conclusive evidence that ties the two clusters together.

In the campaign analyzed by the cybersecurity company, the attack chains have been found to distribute three previously undocumented implants: TernDoor targeting Windows, PeerTime (aka angrypeer) targeting Linux, and BruteEntry, which is installed on network edge devices. The exact initial access method used in the attacks is not known, although the adversary has previously targeted systems running outdated versions of Windows Server and Microsoft Exchange Server to drop web shells for follow-on activity. TernDoor is deployed through DLL side-loading, leveraging the legitimate executable “wsprint.exe” to launch a rogue DLL (“BugSplatRc64.dll”) that decrypts and executes the final payload in memory. A variant of Crowdoor (itself a variant of SparrowDoor), the backdoor is said to have been put to use by UAT-9244 since at least November 2024.

It establishes persistence on the host by means of a scheduled task or the Registry Run key. It also exhibits differences with CrowDoor by making use of a disparate set of command codes and embedding a Windows driver to suspend, resume, and terminate processes. Furthermore, it only supports one command-line switch (“-u”) to uninstall itself from the host and delete all associated artifacts. Once launched, it runs a check to make sure that it has been injected into “msiexec.exe,” after which it decodes a configuration to extract the command-and-control (C2) parameters.

Subsequently, it establishes communication with the C2 server, allowing it to create processes, run arbitrary commands, read/write files, collect system information, and deploy the driver to hide malicious components and manage processes. Further analysis of the UAT-9244’s infrastructure has led to the discovery of a Linux peer-to-peer (P2P) backdoor dubbed PeerTime, which is compiled for several architectures (i.e., ARM, AARCH, PPC, and MIPS) so as to infect a variety of embedded systems. The ELF backdoor, along with an instrumentor binary, is deployed via a shell script. “The instrumentor ELF binary will check for the presence of Docker on the compromised host using the commands docker and docker –q,” Talos researchers Asheer Malhotra and Brandon White said.

“If Docker is found, then the PeerTime loader is executed. The instrumentor consists of debug strings in Simplified Chinese, indicating that it is a custom binary created and deployed by Chinese-speaking threat actors.” The primary goal of the loader is to decrypt and decompress the final PeerTime payload and execute it directly in memory. PeerTime comes in two flavors: one version written in C/C++ and a newer variant programmed in Rust. Besides having the ability to rename itself as a harmless process to sidestep detection, the backdoor employs the BitTorrent protocol to fetch C2 information, download files from its peers, and execute them on the compromised system.

Also staged in the threat actor’s servers are a set of shell scripts and payloads, including a brute-force scanner codenamed BruteEntry that’s installed on edge devices to turn them into mass-scanning proxy nodes within an Operational Relay Box (ORB) capable of brute-forcing Postgres, SSH, and Tomcat servers. This is accomplished by means of a shell script that drops two Golang-based components: an orchestrator that delivers BruteEntry, which then contacts a C2 server to obtain the list of IP addresses to be targeted for performing brute-force attacks. The backdoor ultimately reports successful logins back to the C2 server. “‘Success’ indicates if the brute force was successful (true or false), and ‘notes’ provides specific information on whether the brute force was successful,” Talos said.

“If the login failed, the note reads ‘All credentials tried.’” Found this article interesting? Follow us on Google News , Twitter and LinkedIn to read more exclusive content we post.

Microsoft Reveals ClickFix Campaign Using Windows Terminal to Deploy Lumma Stealer

Microsoft on Thursday disclosed details of a new widespread ClickFix social engineering campaign that has leveraged the Windows Terminal app as a way to activate a sophisticated attack chain and deploy the Lumma Stealer malware. The activity, observed in February 2026, makes use of the terminal emulator program instead of instructing users to launch the Windows Run dialog and paste a command into it. “This campaign instructs targets to use the Windows + X → I shortcut to launch Windows Terminal (wt.exe) directly, guiding users into a privileged command execution environment that blends into legitimate administrative workflows and appears more trustworthy to users,” the Microsoft Threat Intelligence team said in a series of posts on X. What makes the latest variant notable is that it bypasses detections specifically designed to flag Run dialog abuse, not to mention take advantage of the legitimacy of Windows Terminal to trick unsuspecting users into running malicious commands delivered via bogus CAPTCHA pages, troubleshooting prompts, or other verification-style lures.

The post-compromise attack chain is also unique: when the user pastes a hex-encoded, XOR-compressed command copied from the ClickFix lure page into a Windows Terminal session, it spans additional Terminal/PowerShell instances to ultimately invoke a PowerShell process responsible for decoding the script. This, in turn, leads to the download of a ZIP payload and a legitimate but renamed 7-Zip binary, the latter of which is saved to disk with a randomized file name. The utility then proceeds to extract the contents of the ZIP file, triggering a multi-stage attack chain that involves the following steps - Retrieving more payloads Setting up persistence via scheduled tasks Configuring Microsoft Defender exclusions Exfiltrating machine and network data Deploying Lumma Stealer using a technique called QueueUserAPC() by injecting the malware into “chrome.exe” and “msedge.exe” processes “The stealer targets high-value browser artifacts, including Web Data and Login Data, harvesting stored credentials and exfiltrating them to attacker-controlled infrastructure,” Microsoft said. The Windows maker said it also detected a second attack pathway, as part of which, when the compressed command is pasted into Windows Terminal, it downloads a randomly named batch script to the “AppData\Local” folder by means of “cmd.exe” in order to write a Visual Basic Script to the Temp folder (aka %TEMP%).

“The batch script is then executed via cmd.exe with the /launched command-line argument. The same batch script is then executed through MSBuild.exe, resulting in LOLBin abuse,” it added. “The script connects to Crypto Blockchain RPC endpoints, indicating an etherhiding technique. It also performs QueueUserAPC()-based code injection into chrome.exe and msedge.exe processes to harvest Web Data and Login Data.” Found this article interesting?

Hikvision and Rockwell Automation CVSS 9.8 Flaws Added to CISA KEV Catalog

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) on Thursday added two security flaws impacting Hikvision and Rockwell Automation products to its Known Exploited Vulnerabilities ( KEV ) catalog, citing evidence of active exploitation. The critical-severity vulnerabilities are listed below - CVE-2017-7921 (CVSS score: 9.8) - An improper authentication vulnerability affecting multiple Hikvision products that could allow a malicious user to escalate privileges on the system and gain access to sensitive information. CVE-2021-22681 (CVSS score: 9.8) - An insufficiently protected credentials vulnerability affecting multiple Rockwell Automation Studio 5000 Logix Designer, RSLogix 5000, and Logix Controllers that could allow an unauthorized user with network access to the controller to bypass the verification mechanism and authenticate with it, as well as alter its configuration and/or application code.

The addition of CVE-2017-7921 to the KEV catalog comes more than four months after the SANS Internet Storm Center disclosed that it had detected exploit attempts against Hikvision cameras susceptible to the flaw. However, there appears to be no public report describing attacks involving CVE-2021-22681 . In light of active exploitation, Federal Civilian Executive Branch (FCEB) agencies are recommended to update to the latest supported software versions by March 26, 2026, as part of Binding Operational Directive (BOD) 22-01 . “These types of vulnerabilities are frequent attack vectors for malicious cyber actors and pose significant risks to the federal enterprise,” CISA said.

“Although BOD 22-01 only applies to FCEB agencies, CISA strongly urges all organizations to reduce their exposure to cyberattacks by prioritizing timely remediation of KEV Catalog vulnerabilities as part of their vulnerability management practice.” Found this article interesting? Follow us on Google News , Twitter and LinkedIn to read more exclusive content we post.

Cisco Confirms Active Exploitation of Two Catalyst SD-WAN Manager Vulnerabilities

Cisco has disclosed that two more vulnerabilities affecting Catalyst SD-WAN Manager (formerly SD-WAN vManage) have come under active exploitation in the wild. The vulnerabilities in question are listed below - CVE-2026-20122 (CVSS score: 7.1) - An arbitrary file overwrite vulnerability that could allow an authenticated, remote attacker to overwrite arbitrary files on the local file system. Successful exploitation requires the attacker to have valid read-only credentials with API access on the affected system. CVE-2026-20128 (CVSS score: 5.5) - An information disclosure vulnerability that could allow an authenticated, local attacker to gain Data Collection Agent (DCA) user privileges on an affected system.

Successful exploitation requires the attacker to have valid vManage credentials on the affected system. Patches for the security defects, along with CVE-2026-20126, CVE-2026-20129, and CVE-2026-20133, were released by Cisco late last month in the following versions - Earlier than Version 20.91 - Migrate to a fixed release. Version 20.9 - Fixed in 20.9.8.2 Version 20.11 - Fixed in 20.12.6.1 Version 20.12 - Fixed in 20.12.5.3 and 20.12.6.1 Version 20.13 - Fixed in 20.15.4.2 Version 20.14 - Fixed in 20.15.4.2 Version 20.15 - Fixed in 20.15.4.2 Version 20.16 - Fixed in 20.18.2.1 Version 20.18 - Fixed in 20.18.2.1 “In March 2026, the Cisco PSIRT became aware of active exploitation of the vulnerabilities that are described in CVE-2026-20128 and CVE-2026-20122 only,” the networking equipment major said. The company did not elaborate on the scale of the activity and who may be behind it.

Ryan Dewhurst, watchTowr’s head of proactive threat intelligence Ryan Dewhurst, said the preemptive exposure management platform has witnessed exploitation attempts from numerous unique IP addresses and observed threat actors deploying web shells. “The largest spike in activity occurred on March 4, with attacks widely spread across various regions worldwide, and U.S.-based areas saw slightly higher activity than others,” Dewhurst added. “We expect activity to continue as part of the typical long tail of exploitation, as more threat actors become involved. With mass and opportunistic exploitation at play, any exposed system should be considered compromised until proven otherwise.” In light of active exploitation, users are recommended to update to a fixed software release as soon as possible, and take steps to limit access from unsecured networks, secure the appliances behind a firewall, disable HTTP for the Catalyst SD-WAN Manager web UI administrator portal, turn off network services like HTTP and FTP if not required, change the default administrator password, and monitor log traffic for any unexpected traffic to and from systems.

The disclosure comes a week after the company said a critical security flaw in Cisco Catalyst SD-WAN Controller and Catalyst SD-WAN Manager ( CVE-2026-20127 , CVSS score: 10.0) has been exploited by a highly sophisticated cyber threat actor tracked as UAT-8616 to establish persistent footholds into high-value organizations. This week, Cisco also released updates to address two maximum-severity security vulnerabilities in Secure Firewall Management Center ( CVE-2026-20079 and CVE-2026-20131 , CVSS scores: 10.0) that could allow an unauthenticated, remote attacker to bypass authentication and execute arbitrary Java code as root on an affected device. Found this article interesting? Follow us on Google News , Twitter and LinkedIn to read more exclusive content we post.

Preparing for the Quantum Era: Post-Quantum Cryptography Webinar for Security Leaders

Most organizations assume encrypted data is safe. But many attackers are already preparing for a future where today’s encryption can be broken. Instead of trying to decrypt information now, they are collecting encrypted data and storing it so it can be decrypted later using quantum computers. This tactic—known as “harvest now, decrypt later” —means sensitive data transmitted today could become readable years from now once quantum capabilities mature.

Security leaders who want to understand this risk and how to prepare can explore it in detail in the upcoming webinar on Post-Quantum Cryptography best practices , where experts will explain practical ways organizations can begin protecting data before quantum decryption becomes possible. Why Post-Quantum Cryptography Matters Quantum computing is advancing quickly, and most modern encryption algorithms, such as RSA and ECC, will not remain secure forever. For organizations that must keep data confidential for many years—financial records, intellectual property, government communications—waiting is not an option. A practical approach emerging today is hybrid cryptography , which combines traditional encryption with quantum-resistant algorithms like ML-KEM .

This allows organizations to strengthen security without disrupting existing systems. The Future-Ready Security webinar will explain how hybrid cryptography works in real environments and how organizations can begin transitioning to quantum-safe protections. Preparing for the Quantum Era Organizations preparing for quantum threats are focusing on a few key steps: Identify sensitive data that must remain protected long-term Understand where encryption is used across systems Begin adopting hybrid cryptography strategies Maintain visibility into cryptographic algorithms and compliance needs At the same time, security teams must still inspect encrypted traffic and enforce policies across their networks. Modern Zero Trust architectures play an important role in maintaining this control.

These strategies—and how platforms like Zscaler implement them—will be discussed during the live webinar session designed for IT, security, and networking leaders. What You’ll Learn in the Webinar This session will cover: The growing risk of “harvest now, decrypt later” attacks How ML-KEM hybrid encryption helps organizations transition safely How post-quantum traffic inspection enables policy enforcement at scale Best practices for protecting sensitive data in the quantum era Quantum computing will reshape cybersecurity. Organizations that begin preparing early will be better positioned to protect their most critical data. Join the webinar to learn how to build a practical, quantum-ready security strategy before the threat becomes urgent.

Found this article interesting? This article is a contributed piece from one of our valued partners. Follow us on Google News , Twitter and LinkedIn to read more exclusive content we post.

ThreatsDay Bulletin: DDR5 Bot Scalping, Samsung TV Tracking, Reddit Privacy Fine & More

Some weeks in cybersecurity feel routine. This one doesn’t. Several new developments surfaced over the past few days, showing how quickly the threat landscape keeps shifting. Researchers uncovered fresh activity, security teams shared new findings, and a few unexpected moves from major tech companies also drew attention.

Together, these updates offer a useful snapshot of what is happening behind the scenes in the cyber world right now. From new tactics and campaigns to security and policy changes that could affect millions of users, there is a lot unfolding at once. Below is a quick roundup of the most notable stories making headlines this week. Phishing Campaign Deploys Multiple Malware Strains Ukraine Targeted by SHADOWSNIFF, SALATSTEALER, DEAFTICKK Malware The Computer Emergency Response Team of Ukraine (CERT-UA) has warned of a hacking campaign targeting Ukrainian government institutions using phishing emails containing a ZIP archive (or a link to a website vulnerable to cross-site scripting attacks) to distribute SHADOWSNIFF and SALATSTEALER information-stealing malware and a Go backdoor called DEAFTICKK.

The agency attributed the activity to a threat actor tracked as UAC-0252. The development comes as a suspected Russian espionage campaign is targeting Ukraine with two previously undocumented malware strains, BadPaw and MeowMeow , according to ClearSky. While the campaign is likely said to be the work of APT28, the cybersecurity company did not identify the targets of the campaign or say whether the attacks were successful. Fake RMM Service Spreads RAT via Phishing Threat Actor Masquerades as RMM Vendor to Distribute RAT A new malware-as-a-service (MaaS) dubbed TrustConnect (“trustconnectsoftware[.]com”) masqueraded as a legitimate remote monitoring and management (RMM) tool for $300 per month.

It’s assessed that the threat actor behind TrustConnect was also a prominent user of RedLine Stealer . According to email security firm Proofpoint , multiple threat actors have been observed distributing the malware via phishing emails as of January 27, 2026. The emails claim to be event invites or bid proposals, tricking recipients into clicking on links that lead to the download of bogus executables that install TrustConnect RAT. The RAT backdoors users’ machines and gives attackers full mouse and keyboard control, allowing them to record and stream the victim’s screen.

Some campaigns have also been observed delivering legitimate remote access software like ScreenConnect and LogMeIn Resolve alongside TrustConnect between January 31 and February 3, 2026. Customers who purchase the toolkit are granted access to a dashboard to remotely commandeer infected devices and generate branded installers containing the malware. After Proofpoint took steps to disrupt some of the malware’s infrastructure on February 17, 2026, the threat actor resurfaced with a rebranded version of the malware platform called DocConnect. “Disruptions to MaaS operations like RedLine, Lumma Stealer, and Rhadamanthys have created new opportunities for malware creators to fill gaps in the cybercrime market,” Proofpoint said.

“Although TrustConnect only masqueraded as a legitimate RMM, the lures, attack chains, and follow-on payloads (which include RMMs) show overlap with techniques and delivery methods that are frequently observed in RMM campaigns and used by multiple threat actors.” The development comes amid skyrocketing abuse of legitimate RMM software in cyber attacks. Chrome Moves to Two-Week Release Cycle Google Revises Chrome Release Cycle Google has announced that new Chrome iterations will be released every two weeks, moving away from the current four-week release cycle. Since 2021, Google has been shipping major Chrome versions every four weeks, and since 2023, it has been delivering security updates every week for a reduced patch gap and improved quality. “The web platform is constantly advancing, and our goal is to ensure developers and users have immediate access to the latest performance improvements, fixes, and new capabilities,” Google said .

The new release cycle will also apply to beta releases, starting with Chrome 153, which will arrive on September 8, 2026. TPMS Signals Allow Covert Vehicle Tracking Vehicle Tire Pressure Sensors Enable Silent Tracking Researchers at IMDEA Networks Institute have found that Tire Pressure Monitoring System (TPMS) sensors inside each car wheel broadcast unencrypted wireless signals containing persistent identifiers. While the feature is designed for vehicle safety, each sensor transmits a unique ID that does not change, allowing the same car to be recognized again and tracked over time. This, in turn, opens the door to a low-cost monitoring network that uses software-defined radio receivers near roads (at a distance of up to 40m from the car) and parking areas to collect TPMS messages from thousands of vehicles and build profiles of their movements over time.

“Malicious users could deploy passive receivers on large scales and track citizens without their knowledge. The advantage of such a system, over more traditional camera-based ones, is that no direct line-of-sight is needed with the TPMS sensors, and spectrum receivers could be placed in covert or hidden locations, making them harder to spot by victims,” the researchers warned . “Our results show that TPMS transmissions can be used to systematically infer potentially sensitive information such as the presence, type, weight, or driving pattern of the driver.” The disclosure adds to a growing body of research demonstrating how various components fitted into modern vehicles can become unintended conduits for surveillance and exploits. Telegram Emerges as Cybercrime Command Hub Telegram as an Operational Layer for Cybercrime A new analysis from CYFIRMA has pointed out how Telegram’s structure offers threat actors a way to extend their reach globally without the need for specialized tooling, enable frictionless onboarding of buyers and affiliates, support payment options, and facilitate audience growth.

The emergence of the platform has fundamentally changed the way cyber operations are coordinated, monetized, and publicized. “For financially motivated actors, Telegram functions as a scalable storefront and customer support hub,” the company said . “For hacktivists, it serves as a mobilization and propaganda amplifier. For state-aligned operations, it offers a rapid distribution channel for narratives and leaks.

In many cases, telegram complements and increasingly replaces traditional Tor-based ecosystems by removing technical friction while maintaining operational flexibility.” AuraStealer Infrastructure Revealed New AuraStealer Malware Analyzed A new analysis of AuraStealer from Intrinsec has uncovered 48 command-and-control (C2) domain names linked to the stealer’s operations. The threat actor behind the malware has been found to use .shop and .cfd top-level domains, in addition to routing all traffic through Cloudflare as a reverse proxy to conceal the real server. AuraStealer first appeared on underground hacker forums in July 2025, shortly after the disruption of the Lumma Stealer as part of a law enforcement operation. It was advertised by a user named AuraCorp on the XSS forum.

It comes in two subscription packages: $295/month for Basic and $585/month for Advanced. One of the primary mechanisms through which the stealer is distributed is ClickFix . Malvertising Pushes New Atomic Stealer Variant Malvertising Campaign Drops Atomic Stealer A malvertising campaign is using bogus ads on Google Search results pages to redirect users looking for ways to free up macOS storage to fraudulent web pages hosted on Medium, Evernote, and Kimi AI to serve ClickFix -style instructions that drop a new variant of the Atomic Stealer called malext to steal a wide range of data from compromised macOS systems. The campaign uses more than 50 compromised Google Ads accounts that push “over 485 malicious landing pages, ultimately leading to a ClickFix attack that deployed a potentially new version of AMOS Stealer onto infected systems,” security researcher Gi7w0rm said .

Bots Hammer DRAM Pages for DDR5 Inventory Large-Scale Operation Submits Millions of Web Scraping Requests Targeting DRAM Product Pages A large-scale data gathering operation has submitted more than 10 million web scraping requests to hit DRAM product pages on e-commerce sites in an effort to find sellers carrying desirable DRAM stock. The bots have been found to check the stock of specific RAM kits every 6.5 seconds by using a technique called cache busting to ensure they get the most up-to-date information, DataDome said. “These bots aggressively target the entire supply chain, from consumer RAM to B2B industrial memory providers and raw hardware components like DIMM sockets,” the company said . “Scrapers attempt to avoid detection by adding cache-busting parameters to every request and calibrating their speed to stay just below volumetric alarm thresholds.

By rapidly snapping up the limited DDR5 memory inventory for profitable resale, these bots further deplete the consumer supply, effectively boxing out legitimate customers and driving market prices even higher.” Reddit Fined Over Children’s Data Handling U.K. ICO Fines Reddit £14.47M for Children’s Privacy Failures The U.K. Information Commissioner’s Office (ICO) has fined Reddit £14.47 million for unlawfully processing the personal information of children under the age of 13 and for failing to properly check the age of its users, thereby putting them at risk of being exposed to inappropriate and harmful content online. In July 2025, Reddit introduced age assurance measures that include age verification to access mature content and asking users to declare their age when opening an account.

Reddit said it would appeal the decision, stating it doesn’t require users to share information about their identities, regardless of age, to ensure users’ online privacy and safety. Samsung Restricts TV Data Collection in Texas Samsung TVs to Stop Collecting Texans’ Data Without Their Consent Texas Attorney General Ken Paxton announced that Samsung will no longer collect Automated Content Recognition ( ACR ) data without consumers’ express consent. The development comes in the wake of a lawsuit filed against the South Korean electronics giant for its data collection practices and over allegations that the collected ACR information could be used to serve targeted ads. “Additionally, it compels Samsung to promptly update its smart TVs and implement disclosures and consent screens that are clear and conspicuous to ensure that Texans can make an informed decision regarding whether their data is collected and how it’s used,” the Office of the Attorney General said .

Samsung has denied it spies on users. NATO Clears Consumer iPhones and iPads Apple iPhone and iPad Cleared for Classified NATO Use Apple iPhones and iPads have been approved to handle classified information in NATO networks. They are the first consumer-grade devices to be approved for NATO use without additional special software or settings. iPhone and iPad previously received approval to handle classified German government data on devices using native iOS and iPadOS security measures following a security evaluation conducted by Germany’s Federal Office for Information Security.

TikTok Rejects End-to-End Encryption for DMs TikTok Has No Plans to Encrypt Messages ByteDance’s TikTok said it has no plans to add end-to-end encryption (E2EE) to direct messages because it would prevent law enforcement and safety teams from reading messages if necessary. In a statement shared with the BBC, the company said it wanted to protect users, especially young people, from harm. Multi-Stage Phishing Attack Spreads Agent Tesla Phishing Campaign Delivers Agent Tesla A new phishing campaign using purchase order lures has leveraged a multi-stage attack chain to deliver Agent Tesla , allowing threat actors to harvest sensitive data, while taking steps to evade detection using techniques like obfuscation and in-memory execution. “From the initial obfuscated JSE loader to the reflective loading of .NET assemblies and process hollowing of legitimate Windows utilities, Agent Tesla is designed to stay invisible,” Fortinet FortiGuard Labs said .

“Its extensive anti-analysis checks further ensure that it only reveals its true nature when it’s certain it isn’t being watched.” Attackers Abuse Infrastructure-Only .arpa Domain Phishing Attacks Exploit .arpa TLD With organizations taking steps to tighten their traditional email and web filters, new research from Infoblox has found a novel campaign where actors are abusing the .arpa top-level domain, a space strictly reserved for network infrastructure, to host malicious content and bypass standard blocklists. The development shows cybercriminals are finding “impossible” hiding spots within the internet’s core infrastructure to bypass security, the DNS threat intelligence firm said. Elsewhere, threat actors are also abusing LNK shortcut files and WebDAV to download malicious files on targets’ systems. “Because being able to remotely access things on the internet via File Explorer is a relatively unknown functionality to most people, WebDAV is an exploitable way to make people download files without going through a traditional web browser file download,” Cofense said .

Spoofed Email Chains Target LastPass Users LastPass Warns of New Phishing Campaign A new phishing campaign that commenced on March 1, 2026, is using lures related to unauthorized access to individuals’ accounts to trick recipients into visiting fake LastPass login pages to take control of their accounts. The attack takes advantage of the fact that many email clients, especially mobile, show only the display name, hiding the real sender address unless users expand it. “Attackers are forwarding fake email chains to make it appear as though another individual is trying to take unauthorized action on their LastPass account (i.e., export vault, full account recovery, new trusted device registered, etc.),” LastPass said . “Attackers use display name spoofing so that the name portion of the sender field is manipulated to impersonate LastPass, while the actual sending email address is unrelated.” Experts Warn Against Blind Trust in AI Coding Agents Why Relying Fully on an AI Coding Agent Isn’t a Good Idea With the emergence of tools like Claude Code Security , OX Security is urging users to resist the temptation to outsource judgment, architecture, and validation to a single artificial intelligence (AI) model.

“AI doesn’t invent fundamentally new code patterns,” it said . “It reproduces the most common ones it has seen before. That means it scales not only productivity, but also existing weaknesses in software engineering practice.” The cybersecurity company also warned that AI systems may be prone to false positives and may not reliably inform a user if an issue flagged in a single repository is actually exploitable in a complex and unique environment. A pipeline that relies on the same AI system for both writing and reviewing code is not ideal, it added.

LLMs Enable Automated Internet Deanonymization LLMs Are Good at Unmasking Pseudonymous Users at Scale A team of academics from Anthropic, ETH Zurich, and MATS Research has developed large language models (LLMs) that can deanonymize internet users based on past comments or other digital clues they leave behind. “Given two databases of pseudonymous individuals, each containing unstructured text written by or about that individual, we implement a scalable attack pipeline that uses LLMs to: (1) extract identity-relevant features, (2) search for candidate matches via semantic embeddings, and (3) reason over top candidates to verify matches and reduce false positives,” the researchers said . The method works even if targets use different pseudonyms across multiple platforms. The researchers said using their LLMs outperforms classical research methods, where digital footprints are examined manually by a human operator.

This, in turn, enables fully automated deanonymization attacks that can work on unstructured data at scale, while also reducing the cost and effort that goes into intelligence gathering. “Our results show that the practical obscurity protecting pseudonymous users online no longer holds and that threat models for online privacy need to be reconsidered,” the researchers said. “The average online user has long operated under an implicit threat model where they have assumed pseudonymity provides adequate protection because targeted deanonymization would require extensive effort. LLMs invalidate this assumption.” That wraps up this week’s quick look at what has been happening across the cybersecurity landscape.

Each update on its own may seem small, but together they show how quickly things continue to change. New techniques appear, old tactics evolve, and security decisions from major companies can shift the wider ecosystem. For security teams, researchers, and anyone who follows the threat landscape, keeping track of these signals helps make sense of the bigger picture. Stay tuned for the next edition of the ThreatsDay Bulletin with more developments from the cyber world.

Found this article interesting? Follow us on Google News , Twitter and LinkedIn to read more exclusive content we post.