When people discuss the security implications of Unicode, International Domain Names (IDNs) are often highlighted as a risk. However, while visible and often talked about, IDNs are probably not what you should really worry about when it comes to Unicode. There are several issues that impact application security beyond confusing domain names.
ShadowRay 2.0 is an active global cyber campaign that hijacks Ray AI infrastructure to create a self-propagating botnet. The attackers exploit AI systems to turn them against themselves, leveraging the compromised AI nodes to expand the botnet autonomously. This campaign is currently ongoing and has been reported by Oligo Security with limited public technical details. Although no known exploits are confirmed in the wild, the campaign's medium severity rating reflects its potential to disrupt AI-based services and infrastructure. European organizations using Ray AI infrastructure or related AI platforms could face risks of service disruption, data compromise, or being unwitting participants in botnet activities. Mitigations require proactive monitoring of AI infrastructure, network segmentation, and rapid incident response capabilities. Countries with significant AI technology adoption and critical infrastructure relying on AI are more likely to be targeted. Given the campaign's self-propagating nature and AI focus, the threat severity is assessed as high due to potential widespread impact and complexity of mitigation.
The EchoGram flaw is a recently disclosed vulnerability that allows attackers to bypass guardrails implemented in major large language models (LLMs). This bypass enables the LLMs to generate outputs that would normally be restricted by safety and content moderation filters. Although no known exploits are currently in the wild, the flaw poses a medium-level risk due to its potential to facilitate malicious use of LLMs. The vulnerability does not have detailed technical disclosures or patch information yet, limiting immediate mitigation options. European organizations using LLMs for customer interaction, content generation, or automation could face risks related to misinformation, policy violations, or reputational damage. Countries with high adoption of AI technologies and strict regulatory environments are particularly at risk. Mitigation should focus on enhanced monitoring of LLM outputs, restricting sensitive use cases, and collaborating with vendors for updates. Given the medium severity, the threat requires attention but is not immediately critical. Defenders should prioritize awareness and prepare for potential future exploits as more details emerge.
For the latest discoveries in cyber research for the week of 17th November, please download our Threat Intelligence Bulletin. TOP ATTACKS AND BREACHES Cl0p’s Oracle E-Business Suite (CVE-2025-61882) zero-day campaign continues to expand. There are new confirmed breaches at The Washington Post, Logitech, Allianz UK, and GlobalLogic, as well as a newly listed but unconfirmed […] The post 17th November – Threat Intelligence Report appeared first on Check Point Research .
The German Federal Office for Information Security (BSI) has issued guidelines to counter evasion attacks targeting large language models (LLMs). Evasion attacks attempt to manipulate or bypass LLMs' safety and content filtering mechanisms, potentially causing them to generate harmful or unintended outputs. While no known exploits are currently active in the wild, the guidance highlights emerging risks as LLMs become more widely integrated into applications. The threat primarily affects organizations deploying LLM-based systems, especially those handling sensitive or regulated data. European organizations using LLMs must be aware of these evasion techniques to maintain confidentiality, integrity, and compliance. The BSI guidelines provide practical countermeasures to detect and mitigate such attacks. Countries with significant AI adoption and critical infrastructure reliance on LLMs are most likely to be impacted. Given the medium severity and lack of active exploits, proactive mitigation is essential to prevent future exploitation. This threat underscores the evolving security challenges posed by AI technologies in Europe.
The Kraken ransomware group, originating from the HelloKitty cartel, conducts sophisticated big-game hunting and double extortion attacks targeting Windows, Linux, and VMware ESXi systems. They exploit SMB vulnerabilities for initial access and use tools like Cloudflared and SSHFS for persistence and data exfiltration. Kraken ransomware features advanced capabilities including multi-threaded encryption, encryption benchmarking, anti-analysis techniques, and self-deletion to evade detection. The group operates a public data leak site and has launched an underground forum called 'The Last Haven Board' to facilitate their activities. They target a wide range of file types, including SQL databases and network shares, increasing their impact on enterprise environments. Although no known exploits are currently reported in the wild, the threat is medium severity due to its cross-platform nature and complex attack techniques. European organizations, especially those with VMware ESXi and SMB-exposed systems, face significant risk. Denmark is specifically noted as affected, with other European countries likely at risk based on market penetration and strategic targets. Mitigation requires targeted patching of SMB vulnerabilities, network segmentation, monitoring for Cloudflared and SSHFS usage, and robust incident response plans.
A newly reported macOS infection vector leverages AppleScripts to bypass Apple's Gatekeeper security mechanism. Gatekeeper is designed to prevent untrusted or malicious software from executing, but attackers can exploit AppleScript capabilities to circumvent these protections. This technique does not rely on known vulnerabilities but abuses legitimate scripting functionality to execute malicious code. Although no active exploits have been observed in the wild yet, the method poses a medium-level threat due to its potential to evade standard macOS defenses. European organizations using macOS devices, especially in sectors with high reliance on Apple hardware, could be targeted. Mitigation requires enhanced monitoring of AppleScript execution, restricting script permissions, and user education about suspicious script behaviors. Countries with significant macOS adoption and strong tech sectors, such as Germany, the UK, France, and the Nordics, are more likely to be affected. The threat is assessed as medium severity given the moderate impact and the need for some user interaction or social engineering to trigger the script. Defenders should prioritize detection of unusual AppleScript activity and consider application whitelisting policies to reduce risk.
Denmark and Norway are investigating a security flaw found in Yutong buses amid growing concerns about technology vulnerabilities in public transportation. The flaw reportedly affects Yutong buses, a widely used brand in several European countries, raising fears about potential exploitation that could impact passenger safety and operational integrity. Although no known exploits are currently active in the wild, the medium severity rating suggests a tangible risk that requires attention. The investigation highlights the increasing cybersecurity risks in connected vehicle systems, especially in critical public infrastructure. European organizations operating or regulating public transport systems using Yutong buses should be vigilant and prioritize security assessments. Mitigation should focus on thorough vulnerability assessments, patch management, and network segmentation to reduce attack surfaces. Countries with significant Yutong bus deployments and strategic public transport infrastructure, such as Denmark, Norway, Germany, and the UK, are most likely to be affected. Given the potential impact on safety and service availability, the threat severity is assessed as medium. Immediate proactive measures can prevent exploitation and ensure public safety and trust in transportation systems.
ClickFix prompts typically contain instructions for Windows users, but now they are tailored for macOS and they are getting increasingly convincing. The post ClickFix Attacks Against macOS Users Evolving appeared first on SecurityWeek .
An attacker can inject indirect prompts to trick the model into harvesting user data and sending it to the attacker’s account. The post Claude AI APIs Can Be Abused for Data Exfiltration appeared first on SecurityWeek .
A new malware threat leveraging NFC relay attacks targets Android tap-to-pay transactions, enabling attackers to clone payment credentials by intercepting and relaying NFC signals. This malware can effectively bypass typical proximity requirements of NFC payments, allowing fraudulent transactions without physical card or device possession. Although no known exploits are currently active in the wild, the technique poses a medium-level risk due to its potential to compromise payment confidentiality and integrity. European organizations with high adoption of Android tap-to-pay services, especially in retail and financial sectors, are at risk. Mitigation requires enhanced transaction authentication, user awareness, and device security hardening. Countries with widespread NFC payment adoption and significant mobile payment infrastructure, such as the UK, Germany, France, and the Nordics, are most likely affected. The threat is medium severity given the complexity of exploitation and limited current exploitation evidence. Defenders should prioritize monitoring for unusual NFC activity and enforce multi-factor authentication for payment approvals.
The Everest group has claimed responsibility for a cyberattack targeting Svenska kraftnät, Sweden's national power grid operator. While details remain sparse and unverified, the incident raises concerns about the security of critical infrastructure in Sweden. No technical specifics or evidence of exploitation have been publicly disclosed, and there are no known exploits in the wild linked to this claim. The attack, if confirmed, could impact the confidentiality, integrity, and availability of power grid operations, potentially disrupting energy supply. European organizations, especially those involved in critical infrastructure, should remain vigilant. Mitigation should focus on enhanced monitoring, incident response readiness, and securing operational technology environments. Sweden is the primary affected country, with potential indirect effects on neighboring European countries due to interconnected energy networks. Given the lack of detailed information and confirmed impact, the threat severity is assessed as medium. Defenders should prioritize verifying the claim, assessing system integrity, and reinforcing cybersecurity controls in critical infrastructure sectors.
The DNS0. EU private DNS service has shut down due to sustainability issues, impacting users relying on this service for DNS resolution. While this is not a direct security vulnerability or exploit, the sudden unavailability of a DNS service can cause disruptions in network connectivity and potentially increase exposure to DNS-related attacks if users switch to less secure alternatives. European organizations using DNS0. EU may face operational challenges and should ensure they migrate to trusted DNS providers promptly. No known exploits or vulnerabilities are associated with this shutdown. The threat is primarily operational and availability-related rather than a direct cybersecurity compromise. Organizations should review their DNS configurations and monitor for any unusual DNS traffic or redirection attempts during the transition. Countries with higher adoption of DNS0. EU or critical infrastructure relying on it are more likely to be affected.
Attackers are embedding blocks of links that lead to porn sites, online casinos, and the like. We explain their objective, and how to protect your website.
A sophisticated campaign targeting macOS developers has been uncovered, utilizing fake websites impersonating trusted platforms like Homebrew, TradingView, and LogMeIn to distribute Odyssey Stealer and AMOS malware. The attackers employ social engineering tactics, prompting users to paste base64-encoded commands in Terminal, which downloads malicious payloads. Over 85 phishing domains were identified, linked through shared SSL certificates and infrastructure. The campaign's infrastructure includes long-standing IP addresses showing multi-year activity. The malware attempts privilege escalation, performs anti-analysis checks, and disrupts backup services. This coordinated operation demonstrates the attackers' ability to adapt tactics and maintain persistence in the macOS ecosystem.
Stealit malware is a newly identified threat that uses Node.js to conceal itself within fake game and VPN installers. It masquerades as legitimate software to trick users into installation, leveraging the popularity of gaming and VPN tools. Although no known exploits are currently active in the wild, the malware's use of Node.js for obfuscation complicates detection and analysis. The threat primarily targets end users who download software from untrusted sources, potentially leading to data theft or system compromise. European organizations could face risks if employees install these fake applications, especially in sectors with high VPN usage or gaming communities. Mitigation requires enhanced endpoint protection capable of detecting Node.js-based malware, user education on verifying software sources, and network monitoring for unusual outbound connections. Countries with significant gaming markets and high VPN adoption, such as Germany, the UK, France, and the Nordics, are more likely to be affected.
Hackers have stolen approximately 70,000 government ID photos from Discord, a popular communication platform. Discord has publicly acknowledged the breach but dismissed claims that the attackers attempted extortion. The stolen data primarily consists of sensitive government-issued identification images, raising concerns about privacy and potential misuse. There is no evidence of active exploitation or widespread impact beyond the data theft. The incident highlights risks associated with storing sensitive personal information on third-party platforms. European organizations using Discord for official or semi-official communications may face increased exposure to identity theft or social engineering attacks. Mitigation should focus on minimizing sensitive data exposure on such platforms and enhancing monitoring for suspicious activities. Countries with high Discord usage and significant government digital engagement are more likely to be affected. The severity is assessed as medium due to the sensitivity of the data and lack of confirmed active exploitation.
How security posture management for AI can protect against model poisoning, excessive agency, jailbreaking and other LLM risks. The post Will AI-SPM Become the Standard Security Layer for Safe AI Adoption? appeared first on SecurityWeek .
