Threat Intelligence | Breaking Cybersecurity News | The Hacker News

While working on the Transparent Tribe's vibeware research, we have encountered two distinct camps, the optimists and the skeptics. What makes the current dialogue unique is that both sides can be right at the same time. There is, however, a clear operational reason why we encounter "AI attacks" primarily on professional social media feeds rather than within our own telemetry logs. In this article, we analyze the factors explaining why Skynet is not here yet, and how, much like a shark, AI does not need to be innovative to be dangerous. LLM Architecture Bias LLMs are mathematically optimized to predict the most likely outcome, while hacking is the art of identifying the statistical anomaly. LLMs are designed to predict the most statistically probable next token. They are excellent at the average, but poor at the exceptional. A hacker, by contrast, is a practitioner of statistical anomaly, actively seeking the low-pro...

Vulnerability management is the continuous process of identifying, assessing, prioritizing, and addressing security weaknesses across systems, applications, and infrastructure. It extends beyond periodic scanning; it includes validating findings, understanding exposure in real-world environments, and tracking remediation over time. Effective vulnerability management combines asset visibility, vulnerability intelligence, and operational context to determine which flaws present actual risk rather than theoretical exposure. Modern IT environments further complicate the process of vulnerability management. Hybrid IT infrastructure, third-party dependencies, and internet-facing services increase the attack surface while generating large volumes of vulnerability data. Security teams must balance operational constraints, such as out-of-support legacy systems and uptime requirements, with the need to quickly reduce exposure. As a result, vulnerability management is no longer limited to coun...

Telegram entered 2025 under unprecedented pressure. Public scrutiny, regulatory attention, and leadership turmoil forced the platform to do something it had long resisted, enforce at scale. Moderation volumes surged, automation expanded, and millions of channels and groups were removed in a single year. On paper, this looks like a turning point.  In practice, it wasn't the collapse of cyber criminal activity on Telegram; it was an evolution, for sure, but not a collapse.  What we are seeing in 2026 is not a mass exodus from the platform, nor a meaningful decline in threat actor coordination. Instead, Telegram's crackdown has triggered a familiar pattern. Criminal ecosystems adapt faster than platforms can reform. Read the just released Telegram report, by Tal Samra and Or Shichrur for evasion methods, statistics and monitoring recommendations: https://checkpoint.cyberint.com/telegrams-crackdown-criminal-resilience Over 43 Million & Channels Blocked  Tele...

When Shai-Hulud 2.0 hit in late 2025, it was a brutal, expensive wake-up call for DevSecOps teams. It showed that the industry's direction of shifting left, where teams pass security onto developers, wasn't the silver bullet everyone hoped for. Pushing that responsibility was fine in theory, but it crumbled quickly because the foundation it was built on was inherently flimsy. As we move further into 2026, we need a more definitive fix to the structural weakness in the pipelines in light of a potential Shai-Hulud 3.0. A major lesson from 2.0 was that internal CI/CD runners were easily hijacked and turned into attack botnets. Teams need to take that finding and come back with a truly proactive defense. A curated catalog is a way for security teams to control exactly what code and components enter their environment, while still giving engineering teams a fast, secure way to build - it is the key to creating a sustainable solution. More on a curated catalog later. The Anatomy o...

In most security operations centers, CVSS quietly dictates remediation priorities. Dashboards are sorted by severity. "Critical" vulnerabilities float to the top. Quarterly summaries celebrate how many 9.0+ findings were closed. On paper, it looks rational. In practice, it's often wrong. CVSS was designed to standardize how vulnerabilities are scored. Its origins and main purpose have been to measure technical severity, including exploit complexity, required privileges, impact on confidentiality, integrity, and availability. It provides a shared language. But where it has perpetually struggled is measuring context within, like whether the asset is internet-facing, how critical it is to the business, and whether attackers are actively exploiting the vulnerability. And context is where real risk lives. How Abstract Scores Turn Vulnerability Management Into "Severity Theater" A vulnerability scored 9.8 in a non-production environment with no external access may demand immediate atten...

Security teams have never had more telemetry. They have also never been more behind. In 2025, organizations faced an average of 1,968 cyber attacks per week , an 18% YoY increase, and nearly a 70% increase since 2023 . That's not just "more noise." It's a signal that attacker throughput is scaling faster than human response models can. At the same time, the attacker playbook shifted in ways that punish slow cycles. Social engineering moved beyond email into multi-channel, cross-platform operations, including new interaction-led techniques like ClickFix, which manipulates users into executing the attack themselves. ClickFix activity increased by roughly 500% and appeared in nearly half of documented malware campaigns. And while humans remain a primary target, attackers are finding even easier traction in unpatched, unmanaged, and inherited exposures. These gaps give adversaries durable footholds long before exposure remediation is implemented. Couple that with automation, and expo...

Brand impersonation in e-commerce has evolved beyond isolated scam websites into a repeatable, industrialized fraud model operating at global scale . CTM360's latest threat intelligence research analyzes a coordinated campaign—referred to as FraudWear —that demonstrates how attackers are systematically exploiting consumer trust in well-known fashion brands through tens of thousands of fraudulent online stores. Unlike traditional phishing operations, these campaigns do not rely on simple deception or low-effort spoofing. Instead, they replicate the full structure and behavior of legitimate e-commerce platforms , including storefront design, product catalogs, checkout workflows, localized marketing, and payment processing. Each site functions as a disposable asset within a broader, resilient fraud ecosystem. Read the full report here: https://www.ctm360.com/reports/fraudwear-brand-impersonating-online-stores Scale and Targeting Patterns CTM360 identified more than 30,000 malicio...

CTM360 researchers have identified a large-scale fraud campaign involving thousands of fake banking websites that actively target users in the United States and the United Kingdom. Over the past year, more than 11,000 fraudulent bank domains were observed, with 8,000+ in the U.S. and 3,000+ in the UK, all operating without regulatory authorization or physical presence. These are not throwaway phishing pages. They are polished, SEO-optimized platforms that impersonate legitimate financial institutions, regulators, and lending services. Not Your Typical Phishing Fraud What sets this campaign apart is operational maturity. The fake banks offer services such as loans, mortgages, grants, and high-limit credit cards, often promising instant approval or no credit checks. Victims are funneled through simplified onboarding flows, fake KYC processes, and staged "approvals" designed to build trust before monetization. Once engaged, users are pressured to pay activation or processing fees ,...

Open-source intelligence (OSINT) was once a discipline primarily associated with criminal investigations and national intelligence work. Today, it has become a critical pillar in a wide range of corporate and operational processes from internal investigations and fraud detection to KYC, third-party validation, and due-diligence assessments. However, despite this shift in importance, OSINT is still frequently performed in an ad-hoc manner: how data is collected, how evidence is preserved, and operational security mechanisms often depend on individual habits rather than standardised practice. In many cases, investigations are even conducted directly from managed corporate devices, putting both the integrity of the intelligence operation and the wider enterprise network at unnecessary risk. This lack of standardisation introduces operational, security, and compliance risks that many organisations do not fully recognise until something goes wrong. Operational Risk Glazer is a sandboxed...