Blahalla

SAP Releases Patches for Security Flaws Allowing Remote Code Execution

gbhackers.com · 4 months ago

SAP recently released security updates addressing fifteen vulnerabilities across its enterprise software products. The most critical issue is a code injection vulnerability in the SAP Quotation Management Insurance application, identified as CVE-2019-17571, which could allow remote code execution. Another critical flaw, CVE-2026-27685, is an insecure deserialization vulnerability in SAP NetWeaver. Administrators are urged to immediately apply these patches to protect systems and review the SAP Support Portal for a complete list of affected products.

Escape Raises $18 Million to Automate Pentesting

www.securityweek.com · 4 months ago

false

Recent Ivanti Endpoint Manager Flaw Exploited in Attacks

www.securityweek.com · 4 months ago

The US Cybersecurity Agency is urging organizations to immediately patch several vulnerabilities. A high-severity authentication bypass flaw, CVE-2026-1603, exists in Ivanti Endpoint Manager, potentially allowing credential data leakage. Another vulnerability, CVE-2021-22054, is a server-side request forgery issue impacting VMware Workspace One UEM, and a remote code execution flaw, CVE-2025-26399, affects SolarWinds Web Help Desk. Federal agencies are required to patch these vulnerabilities within one week.

CISA: Recently patched Ivanti EPM flaw now actively exploited

www.bleepingcomputer.com · 4 months ago

The Cybersecurity and Infrastructure Security Agency has identified a high-severity vulnerability in Ivanti Endpoint Manager software as actively exploited. The flaw, tracked as CVE-2026-1603, allows attackers to bypass authentication and steal credentials. Federal agencies are required to patch systems within three weeks, and Ivanti reports it has not received any confirmation of exploitation despite a large number of internet-facing instances. Previous Ivanti EPM vulnerabilities, including CVE-2024-13159, CVE-2024-13160, CVE-2024-13161, and CVE-2024-29824, have also been exploited in the wild.

Devs looking for OpenClaw get served a GhostClaw RAT

www.csoonline.com · 4 months ago

Researchers have discovered a malicious npm package, disguised as an OpenClaw installer, that deploys a remote access trojan called GhostClaw. The package, identified as “@openclaw-ai/openclawai”, steals system credentials, browser data, cryptocurrency wallets, SSH keys, and Apple Keychain databases. The malware uses social engineering to harvest passwords and establishes persistence by disguising itself as an npm telemetry service. It also functions as a remote access framework, enabling attackers to route traffic and clone browser sessions.

The Zero-Day Scramble is Avoidable: A Guide to Attack Surface Reduction

thehackernews.com · 4 months ago

Organizations often have more internet-facing exposure than they realize, leading to vulnerabilities like the recent ToolShell vulnerability in Microsoft SharePoint, which attackers exploited before a patch was available. Reducing the attack surface involves three key elements: asset discovery, treating exposure as a risk, and continuous monitoring. Teams should integrate with cloud providers, use subdomain enumeration, and prioritize exposure reduction to minimize risk and avoid reactive scrambling when new vulnerabilities emerge. Daily port scanning is recommended to quickly detect newly exposed services.

SIM Swaps Expose a Critical Flaw in Identity Security

www.securityweek.com · 4 months ago

SIM swap attacks are increasingly common, exploiting vulnerabilities in how organizations verify identity and recover accounts. Criminals manipulate mobile carriers to transfer a victim's phone number, allowing them to intercept authentication codes and access sensitive data like email, banking, and cryptocurrency accounts. Organizations should move away from SMS-based authentication and strengthen account recovery processes with device-bound verification methods and threat detection to prevent these attacks, which now target employees and executives alike. Continuous monitoring and enhanced verification from telecommunications providers are also critical to mitigating this growing threat.

APT28 Uses BEARDSHELL and COVENANT Malware to Spy on Ukrainian Military

thehackernews.com · 4 months ago

Russian state-sponsored hackers, known as APT28, are using new malware to spy on Ukrainian military personnel. The group is employing BEARDSHELL and COVENANT, with COVENANT leveraging the Filen cloud storage service for command and control. Researchers have also identified SLIMAGENT, a keylogger with roots in earlier malware like XAgent, which logs keystrokes and captures screenshots. The malware utilizes a unique obfuscation technique previously seen in the XTunnel tool used during the 2016 Democratic National Committee hack.

Microsoft to enable Windows hotpatch security updates by default

www.bleepingcomputer.com · 4 months ago

false

OpenClaw Advisory Surge Highlights Blind Spot Between GitHub and CVE Vulnerability Tracking

gbhackers.com · 4 months ago

The rapid growth of the OpenClaw AI agent platform has highlighted a disconnect between GitHub’s advisory system and traditional CVE vulnerability tracking. Over 255 GitHub Security Advisories have been published related to issues like command execution and authorization checks, but only a portion currently have CVE identifiers. A request to assign CVEs to 170 OpenClaw advisories was ultimately declined, underscoring a growing backlog of unreviewed GitHub advisories that can leave downstream projects unaware of vulnerabilities. This situation emphasizes a need for improved visibility and potentially adapting security tooling to recognize GitHub Security Advisories as primary vulnerability signals.

Gogs Flaw Could Let Attackers Quietly Overwrite Large File Storage Data

gbhackers.com · 4 months ago

A critical security vulnerability, tracked as CVE-2026-25921, has been found in Gogs, a self-hosted Git service. The flaw allows unauthenticated attackers to silently overwrite Git Large File Storage objects across any repository due to a lack of content verification and repository isolation. Attackers can exploit this to replace legitimate project files with malicious backdoors, potentially compromising software supply chains. Users should immediately update to version 0.14.2 or later to address the vulnerability.

How We Hacked McKinsey's AI Platform

codewall.ai · 4 months ago

Researchers demonstrated a successful attack on McKinsey's internal AI platform, Lilli, exploiting a SQL injection vulnerability. An autonomous agent, lacking credentials or insider knowledge, gained full read and write access to the production database within two hours. The attack exposed sensitive data, including 46.5 million chat messages, user accounts, and decades of proprietary research, highlighting the vulnerability of AI prompt layers. McKinsey has since patched the identified endpoint and taken corrective actions.

Auditing the Gatekeepers: Fuzzing "AI Judges" to Bypass Security Controls

unit42.paloaltonetworks.com · 4 months ago

Researchers have identified a method to bypass AI security gatekeepers, known as AI judges, using a new tool called AdvJudge-Zero. This tool utilizes automated fuzzing to discover stealthy input sequences, like benign formatting symbols, that manipulate the AI's decision-making process and lead to policy violations. These attacks appear as standard formatting, making them difficult to detect with traditional security measures, and have a 99% success rate across various AI models. Organizations can mitigate this risk through adversarial training and products like Cortex AI-SPM.

The OT security time bomb: Why legacy industrial systems are the biggest cyber risk nobody wants to fix

www.csoonline.com · 4 months ago

Legacy operational technology (OT) systems pose a significant cybersecurity risk due to outdated infrastructure like Windows XP and vulnerable protocols such as Modbus. The convergence of IT and OT networks creates pathways for attackers, as demonstrated by incidents like Stuxnet and the Colonial Pipeline ransomware attack. Organizations face challenges including downtime concerns, cultural differences between IT and OT teams, and diffused responsibility for security, delaying necessary upgrades. Addressing this "time bomb" requires prioritized risk assessments, network segmentation, specialized monitoring, and a phased modernization approach.

APT28 hackers deploy customized variant of Covenant open-source tool

www.bleepingcomputer.com · 4 months ago

The APT28 hacker group, also known as Fancy Bear, is employing a customized version of the Covenant post-exploitation framework in ongoing espionage operations targeting Ukraine. Researchers have observed the group using BeardShell and Covenant implants to surveil Ukrainian military personnel, exploiting the CVE-2026-21509 vulnerability in Microsoft Office. BeardShell utilizes the Icedrive cloud service for command-and-control and incorporates obfuscation techniques similar to those used in earlier Xtunnel tools. ESET believes this indicates renewed activity from APT28’s malware development team, enhancing their long-term intelligence gathering capabilities.

Jailbreaking the F-35 Fighter Jet

www.schneier.com · 4 months ago

Several countries are expressing concerns regarding their reliance on the United States for software maintenance of F-35 fighter jets. The Dutch Defense Secretary has suggested the possibility of jailbreaking these aircraft to allow for the installation of third-party software. This action would reduce dependence on US-provided software updates and maintenance. It highlights growing concerns about national security and vendor lock-in.

OpenAI to acquire Promptfoo to strengthen AI agent security testing

www.csoonline.com · 4 months ago

OpenAI plans to acquire Promptfoo, an AI testing startup, to enhance the security of AI agents used in enterprise workflows. Promptfoo’s tools help developers identify vulnerabilities like prompt injection and assess adherence to safety guidelines. The technology will be integrated into OpenAI Frontier and the open-source project will continue to be developed. Experts say this acquisition highlights a broader industry shift toward secure and governed AI systems, with organizations increasingly concerned about AI-related security risks like deepfakes and model manipulation.

Cloudflare Pingora Flaws Enable Request Smuggling and Cache Poisoning Attacks

gbhackers.com · 4 months ago

Cloudflare has disclosed vulnerabilities in its open-source Pingora framework, impacting standalone deployments used as ingress proxies. These flaws, tracked as CVE-2026-2833, CVE-2026-2835, and CVE-2026-2836, allow for request smuggling and cache poisoning attacks. Attackers can bypass security controls, hijack user sessions, and distribute malicious content by exploiting premature protocol upgrades, misinterpreting HTTP/1.0 framing, and insecure default cache keys. Users of Pingora should immediately upgrade to version 0.8.0 and review cache key configurations.

OpenAI to Acquire Promptfoo to Address Vulnerabilities in AI Systems

gbhackers.com · 4 months ago

OpenAI is acquiring Promptfoo, an AI security platform, to enhance the security of its Frontier enterprise platform. The integration will provide automated safety testing, threat mitigation, and secure development workflows, addressing concerns around data security and compliance as AI agents access sensitive corporate data. Promptfoo’s technology is already used by many Fortune 500 companies and will enable businesses to proactively identify and fix vulnerabilities like prompt injections and unauthorized data leaks. Despite the acquisition, OpenAI will continue to support Promptfoo’s open-source tools for the broader cybersecurity community.

Pro-Iran propaganda network gains traction with posts about Epstein

www.washingtonpost.com · 4 months ago

A video circulating online appears to depict Donald Trump and Jeffrey Epstein, but experts suspect it may be disinformation originating from pro-Iran sources. The video, shared widely on X, includes manipulated images and audio designed to appear authentic. Accounts promoting Tehran's narrative are gaining significant visibility by leveraging conspiracy theories surrounding Epstein. The spread of this content highlights the potential for foreign actors to exploit sensitive topics to influence public opinion.

Why access decisions are becoming the weakest link in identity security

www.csoonline.com · 4 months ago

false

SurxRAT Android Malware Uses LLMs for Phishing and Data Theft

gbhackers.com · 4 months ago

A new Android Remote Access Trojan, SurxRAT V5, is being sold through a Telegram-based malware-as-a-service platform. Cybercriminals can purchase licensing tiers, ranging from approximately 200,000 to 500,000 in underground currency, to generate and distribute their own custom Android malware builds for surveillance, data theft, and device control. Recent versions of SurxRAT include a large language model module, potentially enabling automated phishing and adaptive social engineering, and a screen locker module for ransomware-style extortion. The platform has over 1,300 registered users, indicating a rapidly expanding cybercrime ecosystem.

iPhone Hacking Toolkit Tied to Russian Espionage May Have Originated in the U.S.

gbhackers.com · 4 months ago

A sophisticated iPhone hacking toolkit called Coruna, initially developed by U.S. defense contractor L3Harris for government surveillance, has been compromised and spread to Russian spies and Chinese cybercriminals. Former Trenchant employee Peter Williams illicitly sold components of the toolkit, valued at $1.3 million, to Operation Zero, a Russian broker, resulting in a seven-year prison sentence. The toolkit, which targets Apple devices running iOS 13 through 17.2.1, was subsequently used in operations like Operation Triangulation, which targeted Russian diplomats and involved exploit chains Photon and Gallium, and has also been linked to Trickbot ransomware. The leak demonstrates how military cyber weapons can enter the wider threat landscape.

Electric Eye – a Rust/WASM Firefox extension to detect AitM proxies via DOM analysis, TLS fingerprinting and HTTP header inspection

bytearchitect.io · 4 months ago

A new browser extension called Electric Eye is being developed to protect users from Adversary-in-the-Middle phishing attacks, specifically Evilginx. The extension examines various aspects of a webpage, including DNS, HTTP headers, TLS, and the DOM, to identify inconsistencies indicative of malicious activity. Electric Eye uses a risk scoring system based on observed indicators and is designed to be user-friendly. It currently achieves a 100 percent true positive rate against Evilginx and is awaiting approval for distribution.

Your Duolingo Is Still Talking to ByteDance: How Pangle Fingerprints You Across Apps After You Said No

www.buchodi.com · 4 months ago

Researchers discovered the Pangle SDK, owned by ByteDance, collects detailed device information even when users deny ad tracking through Apple's App Tracking Transparency (ATT) framework. This data, including exact storage capacity and boot timestamps, creates a unique device fingerprint across different apps, potentially circumventing ATT's purpose. The analysis of 874 Pangle events from 26 apps revealed that even with a zeroed IDFA, ByteDance can link user activity across apps using these fingerprinting signals. This practice potentially violates Apple's guidelines regarding device identification and user data privacy.

I replaced manual pen tests with automation. Here’s what I learned.

www.csoonline.com · 4 months ago

Traditional manual penetration testing has limitations including cost, time constraints, and a lack of continuous validation. Organizations are increasingly adopting automated penetration testing platforms, such as Pentera and NodeZero, which offer continuous, on-demand simulations and immediate feedback. These platforms reveal vulnerabilities that traditional CVSS severity scores may miss, highlighting actual attack paths and exposing gaps in security controls. Automated testing improves team capabilities, validates detection and response, and provides continuous security validation, though manual testing still holds value for specific scenarios.

Threat Actors Mass-Scan Salesforce Experience Cloud via Modified AuraInspector Tool

thehackernews.com · 4 months ago

Salesforce is reporting increased threat actor activity targeting misconfigured Experience Cloud sites. Attackers are using a modified version of the AuraInspector tool, originally developed by Mandiant, to scan sites and extract data by exploiting overly permissive guest user settings. The activity is focused on customer configuration errors, not vulnerabilities in the Salesforce platform itself, and is linked to a group possibly associated with ShinyHunters. Salesforce recommends customers review and restrict guest user access and monitor logs for suspicious activity.

Signed malware posing as Teams and Zoom apps drops RMM backdoors

gbhackers.com · 4 months ago

Cybercriminals are using phishing campaigns to distribute malware disguised as legitimate applications like Microsoft Teams, Zoom, and Adobe Reader. These malicious files, signed with a compromised Extended Validation certificate from TrustConnect Software, install remote monitoring and management (RMM) backdoors such as ScreenConnect, Tactical RMM, and MeshAgent. The attackers use convincing emails and fake download pages to trick users into installing the RMM tools, which then establish persistent remote access and command-and-control connections. Security experts recommend strict controls for RMM tools, multifactor authentication, and blocking non-approved software signing chains to mitigate this threat.

When AI safety constrains defenders more than attackers

www.csoonline.com · 4 months ago

AI safety measures designed to prevent misuse are inadvertently hindering security professionals while attackers easily bypass these restrictions. Security teams are finding that AI tools block legitimate activities like phishing simulations and penetration testing, while attackers leverage jailbroken models and underground markets to generate increasingly sophisticated attacks, including AI-obfuscated phishing campaigns. This creates a significant defense gap, impacting employee training, research, and vulnerability assessments, and ultimately making organizations less secure. Solutions include authentication for authorized users and purpose-built security tools with controlled environments.

Signal Confirms Sophisticated Phishing Scheme Caused Account Compromises

gbhackers.com · 4 months ago

Signal has confirmed targeted phishing campaigns are causing account takeovers among high-profile individuals like government officials and journalists. The attacks involve manipulating users into sharing SMS verification codes and their Signal PIN through impersonation and psychological tactics, allowing attackers to register accounts on their devices. Signal’s infrastructure and encryption remain secure, with the breaches resulting from social engineering rather than technical vulnerabilities. The company advises users that Signal support will never request verification codes through any communication channel.

Hackers Exploit .arpa and IPv6 Infrastructure to Evade Phishing Defenses

www.neuracybintel.com · 4 months ago

NULL

Chinese APT Campaign Uses Middle East Lures to Target Qatar With PlugX

gbhackers.com · 4 months ago

Chinese state-linked cyber espionage groups are exploiting recent geopolitical tensions in the Middle East to target organizations in Qatar. Attackers rapidly deployed the PlugX backdoor and a previously unseen Rust-based malware loader using lures related to regional conflicts, such as missile strikes in Bahrain and attacks on Gulf energy infrastructure. These campaigns employed DLL hijacking techniques, including abuse of the NVDA screen reader, and often deploy Cobalt Strike for post-exploitation activities. The activity indicates an expanding operational focus for Chinese APT groups in the Gulf region, capitalizing on current events for intelligence gathering.

CISA Flags SolarWinds, Ivanti, and Workspace One Vulnerabilities as Actively Exploited

thehackernews.com · 4 months ago

The Cybersecurity and Infrastructure Security Agency has added three vulnerabilities to its list of actively exploited weaknesses. These include a server-side request forgery flaw in VMware Workspace One UEM, identified as CVE-2021-22054, a deserialization vulnerability in SolarWinds Web Help Desk, CVE-2025-26399, and an authentication bypass in Ivanti Endpoint Manager, CVE-2026-1603. Threat actors are currently exploiting the SolarWinds vulnerability, reportedly linked to the Warlock ransomware group, while the VMware vulnerability has been observed in a coordinated campaign. Federal agencies must apply patches for these flaws by mid-March of 2026.

Anthropic Files Lawsuit Against U.S. Government Over Claude Risk Designation

gbhackers.com · 4 months ago

false

GhostClaw Masquerades as OpenClaw in Bid to Plunder Developer Data

gbhackers.com · 4 months ago

A malicious npm package, @openclaw-ai/openclawai, masquerades as the legitimate OpenClaw CLI to deploy an infostealer and remote access trojan, nicknamed GhostClaw. The package uses social engineering to trick developers into entering their system passwords, which are then used to retrieve and execute a second-stage payload. GhostClaw harvests sensitive data, including SSH keys, cloud credentials, and AI agent configurations, and exfiltrates it via multiple channels. Once installed, it functions as a persistent RAT, enabling attackers to execute commands, clone browser profiles, and steal crypto keys.

Hackers Use Microsoft Teams to Manipulate Employees Into Allowing Remote Access

gbhackers.com · 4 months ago

Cybercriminals are targeting finance and healthcare organizations using a sophisticated attack chain involving Microsoft Teams and email bombing. Attackers impersonate IT support staff to trick employees into granting remote access via the Quick Assist tool, deploying a new backdoor called A0Backdoor. This malware uses DLL sideloading and complex techniques, including a time-based decryption and a hidden character, to evade detection and maintain stealth. The A0Backdoor then employs DNS tunneling to exfiltrate data through trusted public resolvers, disguising malicious communication as routine email traffic.

Dutch Intelligence Warns of Russian-Linked Campaign Targeting Signal and WhatsApp Accounts

www.neuracybintel.com · 4 months ago

NULL

Trojanized Red Alert App Spreads Spyware Through Smishing Campaign Targeting Israeli Users

www.neuracybintel.com · 4 months ago

NULL

U.S. Solar Installations Fell in 2025 as Trump Attacked Clean Energy

www.nytimes.com · 4 months ago

false

Last Week in Security (LWiS) - 2026-03-09

blog.badsectorlabs.com · 4 months ago

Several cybersecurity events occurred recently, including the discovery of a powerful iOS exploit kit capable of targeting devices from 2023, and direct attacks on UAE datacenter services using drones. A significant deal involves Ziff Davis selling Ookla and Downdetector for $1.2 billion. Researchers identified a critical authentication bypass vulnerability in pac4j-jwt, noted as CVE-2026-29000, and a pre-authentication SQL injection vulnerability in FortiClient EMS, CVE-2026-21643. New tools emerged, such as an evasion-as-a-service platform and a Windows malware emulation framework.

Beware! Fake CleanMyMac Website can steal your credential through malicious infostealers.

www.neuracybintel.com · 4 months ago

NULL

Security-Tools für KI-Infrastrukturen – ein Kaufratgeber

www.csoonline.com · 4 months ago

Tools, die die KI-Infrastruktur unter die Lupe nehmen, optimieren und absichern, liegen im Trend. Unser Ratgeber führt Sie zur richtigen AI-Security-Posture-Management-Lösung. Julien Tromeur | shutterstock.com Weil sich Generative-AI-Lösungen branchenübergreifend verbreiten, wächst das Sicherheitsbedürfnis der Anwender. Diesem gerecht zu werden, ist vor allem deshalb eine Challenge, weil die Technologie enormen Einfluss auf die IT-Infrastruktur und die Unternehmensdaten nimmt. Und weil kriminelle Cyberakteure längst erkannt haben, welches Potenzial für sie in diesem Umstand schlummert. Gefragt sind deshalb neue, breit gefächerte Schutz- und Notfallmaßnahmen und Sicherheitssoftware, die spezifisch darauf ausgelegt ist, KI-Infrastrukturen abzusichern. Das hat längst diverse Cybersecurity-Anbieter dazu bewogen, entsprechende Lösungen zu entwickeln. Oder bestehende Produkte mit entsprechenden Features anzureichern. Dieses Wachstumssegment des Security-Markts läuft auch unter der Bezeichnung “AI Security Posture Management”, kurz AI-SPM. In diesem Artikel erfahren Sie: was Security-Lösungen ausmacht, die KI-Infrastrukturen absichern. was AI-SPM-Tools leisten sollten und welche Anbieter und Produkte in diesem Bereich wichtig sind. CSPM, DSPM und AI-SPM AI Security Posture Management fokussiert darauf, die Integrität und Sicherheit von KI- und ML-Systemen zu gewährleisten. Dabei umfasst AI-SPM Strategien, Tools und Techniken, um Daten, Pipelines, Applikationen und Services mit Blick auf ihre Sicherheitslage: zu überwachen, zu bewerten und zu optimieren. Bisher wurden Security Posture Management Tools für zwei separate Bereiche entwickelt: Cloud Security Posture Management (CSPM-) Tools sollen den Cloud-Betrieb allgemein absichern, in erster Linie gegen Fehlkonfigurationen und Missbrauch. Data Security Posture Management (DSPM-) Tools sollen vor Datenlecks und Malware-Infektionen schützen. Das Aufkommen von künstlicher Intelligenz (KI) und Large Language Models (LLMs) hat Bedarf für eine dritte Produktkategorie geschaffen, die gemanagte KI-Cloud-Services und ihre SDKs (beispielsweise Hugging Face Transformer oder Azure Open AI) überwacht und KI-Modellmissbrauch verhindert, AI-SPM. Dass das nötig war, unterstreichen diverse Research-Erkenntnisse, -Beiträge und weitere Ressourcen: Eine Studie des API-Spezialisten Kong (Download gegen Daten) kommt zu dem Ergebnis, dass eine Mehrheit der Befragten Wege gefunden hat, Beschränkungen mit Blick auf die KI-Nutzung zu umgehen. Ein Viertel muss sich erst gar nicht mit so etwas wie Guidelines herumschlagen. Die Non-Profit-Organisation MITRE stellt mit seiner Adversarial Threat Landscape for Artificial Intelligence Systems (ATLAS) eine umfassende Datenbank mit Angriffstaktiken zur Verfügung, die auf “in the wild”-Beobachtungen beruht. Auch das MIT betreibt eine aktive Datenbank, die mehr als 1.600 Risiken in Zusammenhang mit KI-Systemen bereithält. Eine weitere Quelle, um sich mit KI-bezogenen Angriffsmethoden auseinanderzusetzen, bietet das 2023 von OWASP veröffentlichte LLM-Exploit-Ranking (PDF). Die Non-Profit-Organisation hat zudem eine Checkliste für GenAI-Sicherheit veröffentlicht. Sich mit diesen Quellen auseinanderzusetzen, empfiehlt sich, bevor Sie sich für ein Sicherheits-Tool oder -Feature aus dem Bereich AI-SPM entscheiden. Was Security Posture Management für KI leisten sollte Tools im Bereich AI Security Posture Management: bieten im Regelfall agentenlose Konfigurationen, greifen auf Cloud-basierte Modelle zu und belassen Daten auf den vorhandenen Plattformen. Letzteres dient sowohl der Sicherheit als auch dazu, die Verlagerung der damit verbundenen, massiven Datenbestände zu vermeiden. Darüber hinaus spielen bei Security-Tools für KI-Infrastrukturen natürlich auch KI-bezogene Funktionen eine Rolle. Zum Beispiel um große Datenmengen zu klassifizieren, zu tracken und gegen mögliche Missbrauchs- und Angriffsversuche abzusichern. Einige Anbieter haben ihre bestehenden CSPM- oder DSPM-Lösungen um AI-SPM-Features erweitert, inklusive Compliance-Prüfverfahren, Best Practices und Richtlinien, die alle drei Security-Posture-Management-Arten abdecken. Andere offerieren umfassendere Lösungen, die eine Vielzahl KI-bezogener Sicherheitsmaßnahmen beinhalten. Zum Beispiel, um: KI-Pipelines und Workloads schützen, zu erkennen, wenn KI-Modelle sensible Daten referenzieren, Trainingsdaten auf Manipulationen durch Dritte oder externe Applikationen zu überprüfen oder KI-Services und -Plattformen abzusichern. Wichtige AI-SPM-Anbieter Im Folgenden haben wir die AI-SPM-Produkte und -Features neun verschiedener Anbieter für Sie zusammengefasst. Sämtliche Lösungen versprechen, Ihre KI-Infrastruktur abzusichern, verlassen sich dazu jedoch auf unterschiedliche Ansätze. Dabei ist zu beachten, dass es sich um einen Markt handelt, der im Wachstum begriffen ist. Die Produkte sind also noch nicht so umfassend ausgestaltet und integriert, wie sie sein könnten. Zudem arbeiten diverse weitere Sicherheitsanbieter aktiv an ähnlichen Offerings. Cyera.io ist auf Datenklassifizierung spezialisiert und hat eine DSPM-Plattform im Angebot, die um AI-SPM-Features erweitert wird. Die Lösung verspricht beispielsweise Einblicke, auf welche Datenklassen und Data Stores Microsoft-Copilot-Nutzer zugreifen können. LegitSecurity hat sich auf die Fahnen geschrieben, das “AI Visibility Gap” schließen zu wollen. Dazu untersucht die AI-SPM-Plattform KI-Modelle, Code Repositories, kryptografische Secrets und andere KI-bezogene Instanzen. Auf dieser Grundlage entstehen schließlich Risk Scores, um entsprechend priorisieren zu können. Mit dieser Lösung können Sie beispielsweise nachvollziehen, welche User Github Copilot auf der Basis von unsicheren KI-Modellen verwenden. Microsoft stellt AI-Security-Posture-Management-Funktionen im Rahmen einer Preview für sein CSPM-Angebot zur Verfügung. Das fertige Produkt soll Ende 2024 zur Verfügung stehen und zum Einsatz kommen, um GenAI-Applikationen in Multi- oder Hybrid-Cloud-Szenarien abzusichern. Dazu wird zum Beispiel eine GenAI-Softwarestückliste (AI BOM) erfasst. Orca Security verspricht mit seiner Mehrzweck-Sicherheitsplattform unter anderem eine “Ende-zu-Ende”-AI-SPM-Lösung. Diese scannt unter anderem mehr als 50 verschiedene KI-Modellquellen und schlägt Alarm, wenn sie dort, oder in Trainingsdaten-Repositories, sensible Informationen oder Geheimnisse entdeckt. Palo Alto Networks hat Ende 2023 die Übernahme des DSPM-Spezialisten Dig Security abgeschlossen und diesen inzwischen vollständig integriert. Das Ergebnis heißt Prisma Cloud AI-SPM und ermöglicht zum Beispiel Top-Level-Scans der KI-Services von AWS, Google Cloud und Azure. Zudem hat der Sicherheitsanbieter MItte 2025 auch den KI-Sicherheitsanbieter Protect AI übernommen. Securiti.ai verspricht mit seinem Produkt “AI Security & Governance” Schutz für KI-Instanzen. Dieses ermöglicht zum Beispiel KI-Modellrisiken zu bewerten und zu klassifizieren, Compliance-Prüfungen vorzunehmen oder Kontrollmaßnahmen für Daten und KI-Systeme zu etablieren. Varonis hat seine Sicherheitsplattform ebenfalls um “AI Security” erweitert. Das ermöglicht unter anderem, risikobehaftete KI-Fehlkonfigurationen zu erkennen und zu beheben, KI-generierte Inhalte mit Sensibilitäts-Labels zu versehen sowie KI-Workloads oder Datenflüsse zu erkennen, die sensible Informationen beinhalten. Für Microsoft Copilot steht ein eigenes (aufpreispflichtiges) Modul zur Verfügung, demnächst sollen weitere für Salesforce Einstein und Google Gemini folgen. Wiz Security verfügt über einschlägige DSPM- und CSPM-Erfahrungswerte und hat auch eine dedizierte KI-SPM-Lösung im Angebot. Diese verspricht zum Beispiel umfassende Einblicke in KI-Pipelines sowie Detektions-Möglichkeiten für Angriffspfade oder Fehlkonfigurationen. Sie wollen weitere interessante Beiträge rund um das Thema IT-Sicherheit lesen? Unser kostenloser Newsletter liefert Ihnen alles, was Sicherheitsentscheider und -experten wissen sollten, direkt in Ihre Inbox.

Hacker abusing .arpa domain to evade phishing detection, says Infoblox

www.csoonline.com · 4 months ago

A new phishing technique has emerged that combines IPv6 tunneling and abuse of the.arpa domain to bypass security defenses. Hackers are exploiting a vulnerability in DNS record management systems to create A records for reverse DNS names, allowing them to host malicious websites on domains that shouldn't resolve to an IP address. This tactic circumvents security measures that rely on domain reputation and blocklists, and has been used to redirect victims to phishing sites requesting credit card information. Security providers recommend monitoring DNS traffic and examining web traffic origins to detect and prevent this abuse.

Microsoft Teams phishing targets employees with A0Backdoor malware

www.bleepingcomputer.com · 4 months ago

Cybercriminals are targeting employees at financial and healthcare organizations through Microsoft Teams. Attackers initiate contact using social engineering, posing as IT support to gain remote access via Quick Assist. They then deploy a new backdoor called A0Backdoor, utilizing digitally signed MSI installers and DLL sideloading techniques to evade detection. Communication with the command-and-control server is masked within DNS traffic.

Anthropic Sues Department of Defense Over ‘Supply Chain Risk’ Label

www.nytimes.com · 4 months ago

false

Google: Cloud attacks exploit flaws more than weak credentials

www.bleepingcomputer.com · 4 months ago

Hackers are increasingly exploiting newly disclosed software vulnerabilities to breach cloud environments, with the time between vulnerability disclosure and attack shrinking to just days. Exploits now account for nearly half of all intrusions, with remote code execution flaws like those in React2Shell (CVE-2025-55182) and XWiki (CVE-2025-24893) being frequently leveraged. Attackers are also using compromised identities and techniques like phishing to gain access, often targeting cloud platforms for data exfiltration and long-term persistence. North Korean actors have also used fraudulent identities and malicious archives to steal cryptocurrency and gain access to sensitive systems.

Dutch govt warns of Signal, WhatsApp account hijacking attacks

www.bleepingcomputer.com · 4 months ago

Russian state-sponsored hackers are targeting government officials, military personnel, and journalists in a phishing campaign impacting Signal and WhatsApp accounts. Attackers are using social-engineering techniques to trick users into sharing verification codes and Signal PINs, allowing them to take over accounts and monitor messages. Another method involves malicious QR codes that link attacker devices to victim accounts, granting access to messages and enabling impersonation. Users are advised to verify the legitimacy of invitations, links, and QR codes, and regularly review linked devices on their accounts.

White House Cyber Strategy Prioritizes Offense

www.darkreading.com · 4 months ago

The White House released a new cyber strategy prioritizing offensive capabilities and deregulation to assert American dominance in cyberspace. A new executive order aims to disrupt transnational criminal organizations involved in ransomware and other cybercrimes through a dedicated National Coordination Center unit. The strategy emphasizes preemption, targeting adversaries before attacks materialize, and utilizes a range of cyber capabilities, including recent operations against Iran and Venezuela. The document outlines six pillars, including detecting adversaries, streamlining regulations, modernizing government networks, hardening critical infrastructure, ensuring technological leadership, and developing the cybersecurity workforce.

Autonomous agent traffic as an emerging attack surface

www.usevigil.dev · 4 months ago

NULL

Mobile spyware campaign impersonates Israel's Red Alert rocket warning system

www.acronis.com · 4 months ago

A cyber campaign is exploiting geopolitical tensions by distributing a fake version of Israel's Red Alert rocket warning app. Attackers use SMS messages to trick users into installing the malicious application, which steals sensitive data including contacts, SMS messages, location, and device account information. The malware bypasses Android security measures and uses techniques to hide its presence. Users are advised to verify SMS senders, check app permissions, and only install applications from trusted sources.