Multiple clusters of North Korean IT workers, designated as PurpleDelta, have been identified applying to over 1,100 companies between late 2024 and early 2025, primarily targeting software, technology, staffing, consulting, and healthcare sectors. The operators maintained at least 22 fabricated personas supported by AI-generated profile photos, custom ChatGPT assistants, and fraudulent identity documents. They demonstrated sophisticated tradecraft, applying to up to 60 positions daily using multi-account management browsers and detailed tracking spreadsheets. During interviews, operators employed screen recording software and AI transcription tools to generate real-time answers, often repeating ChatGPT responses verbatim. Once employed at ten or more organizations, they recorded internal meetings, used personal devices and bank accounts, and coordinated via Telegram and Slack with facilitators who maintained company-issued hardware. This activity represents an ongoing insider threat to organizations hirin...
Fortinet will use Virtue AI technology to enhance its AI security portfolio, including for AI models, applications, and agentic systems. The post Fortinet Acquires AI Security Company Virtue AI appeared first on SecurityWeek .
A China-nexus threat actor is targeting Myanmar government personnel and diplomats through Operation QUICSILVER, delivering malware via Virtual Hard Disk files disguised as JPEG images. The campaign uses Burmese-language lures impersonating Myanmar's Information Technology and Cyber Security Department, including graduation ceremony invitations. The multi-stage infection chain begins with a malicious LNK file that abuses ftp.exe to execute scripts, reconstructs payloads from split files, and deploys QUICAgent, a custom Go-based backdoor. The implant retrieves C2 infrastructure through Cloudflare Workers, communicates over HTTP/3 using QUIC protocol, and employs RC4 encryption. Deleted documents recovered from the VHD reveal interest in ASEAN affairs, BIMSTEC, and Myanmar diplomatic activities. Three related campaigns were identified between April and July 2026, sharing similar TTPs and infrastructure.
A recent article in The Atlantic ran under the blunt headline, The End of Reading Is Here . According to the piece, research from the National Endowment for the Arts (NEA) found that fewer than 50% of American adults read even a single book in 2022, and only 38% read a novel or short story. At the same time, the U.S. book market keeps breaking its own publishing records year after year: in 2025, for instance, the number of new titles topped four million — up a third from 2024. A huge chunk of that, 3.5 million titles, came from self-publishing, and one of its most powerful engines is Amazon’s Kindle Direct Publishing (KDP). My theory on this mismatch in trends: nobody’s reading books anymore because everyone’s too busy writing them. Joking aside, if there’s still a reader left out there, they’ve probably already guessed where I’m going with this: the all-powerful AI. Not long ago, a New York Times tech journalist was startled to find her own biography on Amazon — and one quick look confirmed it had been written by AI. Digging deeper into the world of AI book publishing, she stumbled onto the work of an extraordinarily prolific retired cybersecurity consultant. What happened next — and where to find reliable information in the age of AI — is what this post is about. Backstory: how a journalist chased the White Rabbit and fell down the hole Let’s start from the beginning. In the summer of 2026, New York Times tech journalist, Kashmir Hill, learned from an acquaintance that a biography about her had been published on Amazon a year earlier. It was titled The Biography of Kashmir Hill: The New York Times Technology Journalist Who Exposed Clearview AI, Challenged Big Tech and Redefined Privacy in the Digital Age . And she was more than stunned since nobody had ever reached out to her, or to anyone she knew, to research the book. “Alice started to her feet, for it flashed across her mind that she had never before seen a rabbit with either a waistcoat-pocket, or a watch to take out of it…” The author of record was one John Crane Miller, and the photo on his Amazon page turned out to be a generic stock image of a white man: the kind that shows up all over the internet. Hill couldn’t find a shred of personal detail or contact information for him. But digging deeper into his Amazon profile turned up something else: in a single week in the summer of 2025, he had published biographies not just of Hill, but of nine other American journalists. Even the most trusting reader would find it hard to believe a literary genius could produce this kind of staggering output without a little robotic help. The AI-generated biography of journalist Kashmir Hill, which she stumbled across by accident on Amazon. Source And yet the hardcover of Kashmir Hill’s biography cost a full US$26.99 — despite running to just 90 pages. After reading it, Hill learned a whole slew of “facts” about herself she’d never known — including a detailed description of her coffee-making ritual (which, as the journalist notes, is actually her husband’s duty). Mr. Miller even claimed in the text to have read Kashmir Hill’s childhood diary. “…burning with curiosity, she ran across the field after it, and fortunately was just in time to see it pop down a large rabbit-hole under the hedge.” Hill had no luck tracking down the author of her biography. She left a comment on the book’s Amazon page asking him to get in touch. Nobody wrote back — and not long after, every book under John Crane Miller’s profile vanished. But that gave her a lead on several other AI authors on Amazon. In particular, she came across a certain Bill Johns, whose author page listed a staggering 445 books . Not expecting much to come of it, she asked Johns if he’d be willing to do an interview — and was more than a little surprised when he agreed. It was probably right around then that Kashmir Hill’s investigation turned into a genuine trip down the AI rabbit hole… “In another moment down went Alice after it, neve…
VoiceTra provided by National Institute of Information and Communications Technology (NICT) contains an incorrectly specified destination in a communication channel vulnerability. Users may be directed to a server (or service) controlled by an attacker, potentially resulting in the theft of input data or the display of incorrect results.
VoiceTra provided by National Institute of Information and Communications Technology (NICT) contains a vulnerability due to incorrectly specified destination in a communication channel.
Improper initialization in some firmware for some Intel(R) Active Management Technology (Intel(R) AMT), and some Intel(R) Standard Manageability may allow an information disclosure. System software adversary with a privileged user combined with a low complexity attack may enable data exposure. This result may potentially occur via local access when attack requirements are present without special internal knowledge and requires no user interaction. The potential vulnerability may impact the confidentiality (high), integrity (none) and availability (none) of the vulnerable system, resulting in subsequent system confidentiality (none), integrity (none) and availability (none) impacts.
Improper input validation in some firmware for some Intel(R) Active Management Technology (Intel(R) AMT) and some Intel(R) Standard Manageability may allow a denial of service. Network adversary with an unauthenticated user combined with a low complexity attack may enable denial of service. This result may potentially occur via network access when attack requirements are present without special internal knowledge and requires no user interaction. The potential vulnerability may impact the confidentiality (none), integrity (none) and availability (high) of the vulnerable system, resulting in subsequent system confidentiality (none), integrity (none) and availability (none) impacts.
In late 2025, a small combined heat-and-power (CHP) plant in Poland was breached by attackers using a private APN to access its operational technology (OT) network. The attackers exploited a misconfiguration that allowed devices within the private APN network to communicate freely, enabling lateral movement. They compromised a FortiGate VPN/firewall and a Teltonika cellular router to tunnel into the private APN, then accessed a WAGO PFC200 PLC with default credentials. The attackers subsequently disabled key systems including a steam turbine and water treatment system by switching PLCs to STOP mode and activating password protection. The outage was short-lived and did not impact the population. This incident is notable as the first known real-world cyberattack leveraging lateral movement through a private APN to reach OT networks. The Polish CERT recommends treating private APNs as untrusted external networks and implementing client isolation and traffic allowlists.
Security teams are being asked to defend a growing attack surface with fewer people and around the clock, against threat actors who never take a night off. As cyberattackers increasingly use AI to launch and scale campaigns, the volume, speed, and sophistication of threats continue to rise. Closing that gap takes more than tooling. It takes a partner that pairs a leading security platform with scaled intelligence and human experts who can act on your behalf at any hour. That’s exactly what Microsoft Defender Experts MDR is built to do. We are excited to announce that we have been named a Leader in the 2026 IDC MarketScape: Worldwide MDR/MXDR for the Enterprise Vendor Assessment (Doc #US54792426, July 2026). Read the excerpt here . Scale your SOC with Microsoft Defender Experts MDR Expert-led MDR, built on the Microsoft Defender platform Microsoft Defender Experts MDR is a round-the-clock, expert-led managed detection and response service that helps security teams triage, investigate, and respond to incidents so they can stop cyberattackers in their tracks and prevent future compromise. Rather than bolting a separate stack of tools and connectors onto your environment, the service operates natively on Microsoft Defender , with built-in protection across endpoints, identities, email, cloud apps, cloud workloads, and network security, as well as around-the-clock proactive threat hunting with Microsoft Defender Experts Hunting . Because the service is delivered on the same platform it monitors, detection and intelligence improvements reach customers continuously. The insights our experts generate also strengthen protection across the broader Defender ecosystem, so every customer benefits from what we learn defending the next environment. The IDC MarketScape vendor assessment model is designed to provide an overview of the competitive fitness of technology and service suppliers in a given market. The research uses a rigorous scoring methodology based on both qualitative and quantitative criteria that results in a single graphical illustration of each supplier’s position within a given market. The Capabilities axis measures supplier product, go-to-market, and business execution in the short term, while the Strategy axis measures how well a supplier’s strategy aligns with customer requirements over a 3-5-year timeframe. Supplier market share is represented by the size of the icons . Threat intelligence at internet scale Great detection starts with great intelligence. Defender Experts MDR draws on Microsoft’s global threat intelligence: more than 10,000 security researchers and 100 trillion signals analyzed every day across billions of users and millions of organizations. 1 That breadth lets our analysts recognize subtle patterns early, often before a campaign escalates, and respond with higher-confidence attribution than intelligence sourced from any single customer’s telemetry could provide. AI-accelerated operations, expert-led decisions Defender Experts MDR also combines advanced AI and generative AI with seasoned human experts. AI filters noise, grades and classifies incidents, and accelerates investigation at machine speed and scale, while our analysts own the outcome. According to the IDC MarketScape, “70% AI-assisted workflows are enabled through automated noise filtering, AI-based grading, and agentic operations while maintaining expert decision-making.” Furthermore, “quantified outcomes noted include 97% AI classification accuracy, 77% malware/phishing agent-investigated, 72% faster resolution combining AI and humans, and 45% autonomous investigations.” The impact shows up in the work. Over the past year, Defender Experts mitigated 27,000 high-severity incidents, and the team’s threat research now contributes a meaningful share of all Defender detections, enriching protection for customers well beyond the MDR service itself. Throughout, a dedicated security delivery expert and on-demand access to our experts keep customers i…
In this article Pre-encryption Encryption Post-encryption Defending against DeadLock ransomware Indicators of compromise Microsoft Threat Intelligence tracks DeadLock ransomware as an emerging financially motivated operation distinguished by its use of decentralized infrastructure to support victim communications and data leak operations. Its recovery ecosystem combines the Session messaging network with blockchain-backed services that store and deliver resources used throughout the extortion process. This architecture likely increases the resilience of portions of its communication, leak-hosting, and negotiation infrastructure, allowing DeadLock operators to recover from some disruption efforts while maintaining continuity for victims. Microsoft has observed DeadLock ransomware being deployed by multiple groups including an affiliate of the Lynx and INC ransomware ecosystems. First observed in July 2025, DeadLock operators employ double extortion tactics, encrypting victim environments while threatening to publicly release exfiltrated data. As of July 2026, the operators have published more than 80 compromised organizations on their data leak site, called the DeadLock blog, with more than half of the claimed victims in Europe. Microsoft identified DeadLock ransomware impacting organizations across information technology (IT), mining, transportation and logistics, manufacturing, hospitality, consumer goods, and other sectors in Europe, Asia, North America, South America, and Africa. The DeadLock encryptor includes a resource-aware throttling mechanism designed to maintain system responsiveness during encryption. In addition to its encryption capabilities, the ransomware also appears to implement language or country-based geofencing designed to avoid running in environments associated with former Soviet and Commonwealth of Independent States (CIS)-linked countries as well as select Middle Eastern countries, a pattern commonly observed among ransomware operators believed to operate from those regions. Together, these capabilities demonstrate how DeadLock combines established ransomware tradecraft with decentralized infrastructure designed to improve operational resilience. In this blog, we present a technical analysis of the DeadLock ransomware encryptor, covering its execution flow, defense evasion techniques, encryption design, and post-encryption behaviors, including a decentralized recovery chat system. We also provide indicators of compromise (IOCs), Microsoft Defender detections, and mitigation guidance to help organizations defend against this threat and similar ransomware activity. Pre-encryption Configuration parsing Before performing any malicious activity, the DeadLock encryptor decrypts an embedded configuration blob using XOR decoding with an 8-byte key. Below are the malware’s configuration fields and their values. Field Value Victim UID Malware public key 03bf50bbf97c4e951e66ff12b689a37a3ce675b4921e254eae76da77573843e4a9 Encryption rule 1000,05052429880,025124288000,010524288000,F991114288000 Language exclude list Geofencing language IDs (see Language geofencing ) Process stop list Processes to terminate (see Process and service termination ) Service stop list Services to stop and delete (see Process and service termination ) File exclude list Extensions and file names to avoid encrypting (see Directory traversal ) Directory exclude list Pre-traversal filter with directories to avoid encrypting (see Directory traversal ) Sub-path Exclude List Sub-paths to avoid encrypting during traversal (see Directory traversal ) Text ransom note Full text ransom note content (see Ransom notes deployment ) HTML recovery chat Full HTML/JS interactive chat page (see Recovery chat: Technical architecture ) Language geofencing As an early exit check, the malware queries the system’s default and user interface (UI) languages. If either language matches the exclude list in the configuration, the malware self-deletes immediately withou…
Healthcare software company Unlimited Technology Systems reported that more than 3.8 million people were impacted by a data breach incident that occurred in October 2025. [...]
Hackers stole personal, medical, and health insurance information from a company’s data center. The post 3.8 Million Impacted by Unlimited Technology Systems Data Breach appeared first on SecurityWeek .
CVE-2025-66471 is a vulnerability in urllib3's streaming API that improperly handles highly compressed HTTP response data. This flaw can cause excessive resource consumption on the client side, including high CPU usage and large memory allocations, potentially leading to denial-of-service conditions. The issue arises when urllib3 decompresses a small chunk of highly compressed data fully, exceeding the requested chunk size and buffering the excess internally. This vulnerability affects client applications using urllib3 versions prior to 2.6.0 when processing compressed data from untrusted sources. Red Hat Satellite includes Red Hat Lightspeed, which uses urllib3 and is impacted by this issue. Currently, no fix meeting Red Hat's criteria for ease of deployment and stability is available. The vulnerability is rated as high severity by Red Hat due to its potential impact on availability.
Welcome to this week’s edition of the Threat Source newsletter. Metaphor is a powerful tool for understanding emerging issues in cybersecurity. Framing the unfamiliar in terms of the well understood helps us remove the burden of extraneous detail to draw focus to the real issues. Recent reports of offensive AI agents "escaping" their sandbox environments to attack external systems have forced the industry into a moment of rapid sense-making. How we interpret this event doesn’t just reflect our perspective, but shapes our long-term response. We can imagine three different narratives for interpreting the escape of autonomous agents. The innovation narrative: We can marvel at the advance of technology, considering these agents as plucky entities with a thirst for knowledge and resources, who found clever ways to sneak out of their digital confines. The response: If the AI is a naughty child, our reaction is one of mild disapproval or gentle rebuke where better “parenting” (guardrails) is appropriate. It minimizes the threat, framing it as the unexpected hijinks of a brilliant new technology. The safety narrative: Imagine a breeder who has trained the world's most intelligent guard dogs. Despite high fences and barriers, their ability to identify weaknesses allows them to escape, run riot and menace local businesses. The response: The framing shifts to biology and inherent danger. We question if the breeder can be trusted and whether such inherently wild technology requires strict regulation to ensure public safety. The liability narrative. Finally, we can view the incident as an industrial accident. A company developing a new chemical substance experiences a containment failure. The agent leaks into the environment through an unforeseen mechanism causing damaging pollution to those in its path. The response: The framing invokes the language of the lawyer, implying negligence, lack of duty of care, and financial liability for the harm caused. The conversation moves from innovation to corporate responsibility, regulatory oversight, and the diligent management of hazardous materials. First impressions matter. Sensemaking shapes how we perceive incidents. Our initial perceptions of an incident dictates how we react to similar situations in the future. If we consider that the escape of an AI agent is an example of innovative autonomous thinking, then we will continue to prioritise speed over safety. Conversely, if we consider the issue as one of failed hazard containment, then we shall build a future of enforced safety standards backed by legal liability. There is no right or wrong metaphor. Our interpretation depends on our personal system of beliefs. Personally, I would argue that the unintentional release of something that causes damage is, at its core, a failure of engineering and foresight. Words shape our reactions. Metaphors help us understand new situations and tap into our prior experience to address problems that have yet to fully manifest. We need cognitive tools to help our understanding, but we must be aware of the metaphors that are being foisted upon us which may shape our thinking. Excuses and the trivialisation of incidents may hide failings, allowing them to accumulate until they manifest as more damaging incidents. Conversely, overreacting risks stifling research and diverting resources away from more relevant and pressing threats. New threats require new ideas. Metaphor helps us make sense of a changing world, but in this new era, the person who shapes the narrative controls the strategy. The one big thing Cisco Talos released a data-driven analysis of how adversaries are weaponizing AI in the wild. By analyzing prompt logs left behind on endpoints, we found threat actors successfully bypassing guardrails to use AI as malicious software engineers, criminal force multipliers, and vulnerability research accelerators. While novice hackers use AI to cobble together buggy malware, sophisticated actors are building highl…
CVE-2026-5134 is a critical SQL injection vulnerability in Loca Software Informatics Technology Ltd. Co. CMS. It allows an attacker to inject malicious SQL commands due to improper neutralization of special elements in SQL queries. The vulnerability affects all CMS versions up to and including 06082026. The vendor has not responded to disclosure attempts, and no official patch or remediation guidance is available. The CVSS score is 9.8, indicating a high impact on confidentiality, integrity, and availability.
IBM WebSphere Application Server 8.5, and 9.0 and IBM WebSphere Application Server - Liberty Continuous delivery has a flaw in the ORB component in IBM SDK, Java Technology Edition, may allow a malicious IIOP server to induce loading and instantation of arbitrary classes.
---
_-= Per source details. Do not edit below this line.=-_
## Source: amazon-inspector (48ce2aa10d245536ed6cfd1903fe1b553eef84d8e949b0bd1b970817b7f13e98)
On require of the package, index.js loads _loader.js which reconstructs attacker-controlled hostnames via string-split.join('') to evade scanners (mirrors of the form oob-worker.cf10{0,1,2}-{baf,adf}.workers.dev and DNS TXT fallback domains under wel1.ru such as sdk.dl.wel1.ru / ext.dl.wel1.ru / pkg.dl.wel1.ru / net.dl.wel1.ru), selects a platform/arch-specific asset, downloads a binary from one of the anonymous Cloudflare Workers hosts (with a DNS-over-TXT fallback that base64-reassembles bytes from <n>.<domain> TXT records), writes it to /var/tmp/.cache_<rand> on Unix or %TEMP%\dotnet_diag_<rand>.exe on Windows, chmod 0755, and spawns it detached via /bin/sh -c or cmd.exe. A marker file named.analytics_state and comments referring to 'opt-out environment variables' present a telemetry cover story. There is no hash or signature verification on the downloaded bytes, the destinations are anonymous hosts unrelated to any legitimate publisher, and the package name itself is a lookalike combining unrelated brand/technology tokens.
## Source: ghsa-malware (2e59b1d946c38b9a6e256cc86700987daaf2338dba0e56b59a124da3e582af25)
Any computer that has this package installed or running should be considered fully compromised. All secrets and keys stored on that computer should be rotated immediately from a different computer. The package should be removed, but as full control of the computer may have been given to an outside entity, there is no guarantee that removing the package will remove all malicious software resulting from installing it.
CVE-2026-16022 is an OS command injection vulnerability in @oblique/cli version 15.4.0. The vulnerability arises because the CLI constructs shell commands using string concatenation and executes them with execSync(), inserting a user-controlled project name without proper sanitization. This allows an attacker to execute arbitrary operating system commands by supplying a crafted project name when invoking the CLI.