Vulnerability | Breaking Cybersecurity News | The Hacker News

Cybersecurity researchers have disclosed six security flaws in the Ollama artificial intelligence (AI) framework that could be exploited by a malicious actor to perform various actions, including denial-of-service, model poisoning, and model theft. "Collectively, the vulnerabilities could allow an attacker to carry out a wide-range of malicious actions with a single HTTP request, including denial-of-service (DoS) attacks, model poisoning, model theft, and more," Oligo Security researcher Avi Lumelsky said in a report published last week. Ollama is an open-source application that allows users to deploy and operate large language models (LLMs) locally on Windows, Linux, and macOS devices. Its project repository on GitHub has been forked 7,600 times to date. A brief description of the six vulnerabilities is below - CVE-2024-39719 (CVSS score: 7.5) - A vulnerability that an attacker can exploit using /api/create an endpoint to determine the existence of a file in the se...

Google said it discovered a zero-day vulnerability in the SQLite open-source database engine using its large language model (LLM) assisted framework called Big Sleep (formerly Project Naptime). The tech giant described the development as the "first real-world vulnerability" uncovered using the artificial intelligence (AI) agent. "We believe this is the first public example of an AI agent finding a previously unknown exploitable memory-safety issue in widely used real-world software," the Big Sleep team said in a blog post shared with The Hacker News. The vulnerability in question is a stack buffer underflow in SQLite, which occurs when a piece of software references a memory location prior to the beginning of the memory buffer, thereby resulting in a crash or arbitrary code execution. "This typically occurs when a pointer or its index is decremented to a position before the buffer, when pointer arithmetic results in a position before the beginning of t...

As SaaS providers race to integrate AI into their product offerings to stay competitive and relevant, a new challenge has emerged in the world of AI: shadow AI.  Shadow AI refers to the unauthorized use of AI tools and copilots at organizations. For example, a developer using ChatGPT to assist with writing code, a salesperson downloading an AI-powered meeting transcription tool, or a customer support person using Agentic AI to automate tasks – without going through the proper channels. When these tools are used without IT or the Security team's knowledge, they often lack sufficient security controls, putting company data at risk. Shadow AI Detection Challenges Because shadow AI tools often embed themselves in approved business applications via AI assistants, copilots, and agents they are even more tricky to discover than traditional shadow IT. While traditional shadow apps can be identified through network monitoring methodologies that scan for unauthorized connections based on...

Cybersecurity researchers have flagged a "massive" campaign that targets exposed Git configurations to siphon credentials, clone private repositories, and even extract cloud credentials from the source code. The activity, codenamed EMERALDWHALE , is estimated to have collected over 10,000 private repositories and stored in an Amazon S3 storage bucket belonging to a prior victim. The bucket, consisting of no less than 15,000 stolen credentials, has since been taken down by Amazon. "The stolen credentials belong to Cloud Service Providers (CSPs), Email providers, and other services," Sysdig said in a report. "Phishing and spam seem to be the primary goal of stealing the credentials." The multi-faceted criminal operation, while not sophisticated, has been found to leverage an arsenal of private tools to steal credentials as well as scrape Git config files, Laravel .env files, and raw web data. It has not been attributed to any known threat actor or grou...

Microsoft has revealed that a Chinese threat actor it tracks as Storm-0940 is leveraging a botnet called Quad7 to orchestrate highly evasive password spray attacks. The tech giant has given the botnet the name CovertNetwork-1658, stating the password spray operations are used to steal credentials from multiple Microsoft customers. "Active since at least 2021, Storm-0940 obtains initial access through password spray and brute-force attacks, or by exploiting or misusing network edge applications and services," the Microsoft Threat Intelligence team said . "Storm-0940 is known to target organizations in North America and Europe, including think tanks, government organizations, non-governmental organizations, law firms, defense industrial base, and others." Quad7, aka 7777 or xlogin, has been the subject of extensive analyses by Sekoia and Team Cymru in recent months. The botnet malware has been observed targeting several brands of SOHO routers and VPN appliances...

A high-severity security flaw has been disclosed in the LiteSpeed Cache plugin for WordPress that could allow an unauthenticated threat actor to elevate their privileges and perform malicious actions. The vulnerability, tracked as CVE-2024-50550 (CVSS score: 8.1), has been addressed in version 6.5.2 of the plugin. "The plugin suffers from an unauthenticated privilege escalation vulnerability which allows any unauthenticated visitor to gain administrator level access after which malicious plugins could be uploaded and installed," Patchstack security researcher Rafie Muhammad said in an analysis. LiteSpeed Cache is a popular site acceleration plugin for WordPress that, as the name implies, comes with advanced caching functionality and optimization features. It's installed on over six million sites. The newly identified issue, per Patchstack, is rooted in a function named is_role_simulation and is similar to an earlier flaw that was publicly documented back in August ...

A now-patched security flaw in the Opera web browser could have enabled a malicious extension to gain unauthorized, full access to private APIs. The attack, codenamed CrossBarking , could have made it possible to conduct actions such as capturing screenshots, modifying browser settings, and account hijacking, Guardio Labs said. To demonstrate the issue, the company said it managed to publish a seemingly harmless browser extension to the Chrome Web Store that could then exploit the flaw when installed on Opera, making it an instance of a cross-browser-store attack. "This case study not only highlights the perennial clash between productivity and security but also provides a fascinating glimpse into the tactics used by modern threat actors operating just below the radar," Nati Tal, head of Guardio Labs, said in a report shared with The Hacker News. The issue has been addressed by Opera as of September 24, 2024, following responsible disclosure. That said, this is not th...

Navigating the complexities of compliance frameworks like ISO 27001, SOC 2, or GDPR can be daunting. Luckily, Intruder simplifies the process by helping you address the key vulnerability management criteria these frameworks demand, making your compliance journey much smoother. Read on to understand how to meet the requirements of each framework to keep your customer data safe. How Intruder supports your compliance goals Intruder's continuous vulnerability scanning and automated reporting help you meet the security requirements of multiple frameworks, including SOC 2, ISO 27001, HIPAA, Cyber Essentials, and GDPR. Here are three core ways Intruder can support you: 1. Making vulnerability management easy Security can be complicated, but your tools shouldn't be. Intruder's always-on platform brings together multiple powerful scanning engines, delivering comprehensive protection that goes beyond traditional vulnerability management. Covering application, cloud, internal, and netwo...

A little over three dozen security vulnerabilities have been disclosed in various open-source artificial intelligence (AI) and machine learning (ML) models, some of which could lead to remote code execution and information theft. The flaws, identified in tools like ChuanhuChatGPT, Lunary, and LocalAI, have been reported as part of Protect AI's Huntr bug bounty platform. The most severe of the flaws are two shortcomings impacting Lunary, a production toolkit for large language models (LLMs) - CVE-2024-7474 (CVSS score: 9.1) - An Insecure Direct Object Reference (IDOR) vulnerability that could allow an authenticated user to view or delete external users, resulting in unauthorized data access and potential data loss CVE-2024-7475 (CVSS score: 9.1) - An improper access control vulnerability that allows an attacker to update the SAML configuration, thereby making it possible to log in as an unauthorized user and access sensitive information Also discovered in Lunary is anot...

Sherlock Holmes is famous for his incredible ability to sort through mounds of information; he removes the irrelevant and exposes the hidden truth. His philosophy is plain yet brilliant: "When you have eliminated the impossible, whatever remains, however improbable, must be the truth." Rather than following every lead, Holmes focuses on the details that are needed to move him to the solution. In cybersecurity, exposure validation mirrors Holmes' approach: Security teams are usually presented with an overwhelming list of vulnerabilities, yet not every vulnerability presents a real threat. Just as Holmes discards irrelevant clues, security teams must eliminate exposures that are unlikely to be exploited or do not pose significant risks. Exposure validation (sometimes called Adversarial Exposure Validation) enables teams to concentrate on the most significant issues and minimize distractions. Similar to Holmes' deductive reasoning, validation of exposures directs organizations towa...

More than six years after the Spectre security flaw impacting modern CPU processors came to light, new research has found that the latest AMD and Intel processors are still susceptible to speculative execution attacks. The attack, disclosed by ETH Zürich researchers Johannes Wikner and Kaveh Razavi, aims to undermine the Indirect Branch Predictor Barrier ( IBPB ) on x86 chips, a crucial mitigation against speculative execution attacks. Speculative execution refers to a performance optimization feature wherein modern CPUs execute certain instructions out-of-order by predicting the branch a program will take beforehand, thus speeding up the task if the speculatively used value was correct. If it results in a misprediction, the instructions, called transient, are declared invalid and squashed, before the processor can resume execution with the correct value. While the execution results of transient instructions are not committed to the architectural program state, it's still ...

A new attack technique could be used to bypass Microsoft's Driver Signature Enforcement (DSE) on fully patched Windows systems, leading to operating system (OS) downgrade attacks. "This bypass allows loading unsigned kernel drivers, enabling attackers to deploy custom rootkits that can neutralize security controls, hide processes and network activity, maintain stealth, and much more," SafeBreach researcher Alon Leviev said in a report shared with The Hacker News. The latest findings build on an earlier analysis that uncovered two privilege escalation flaws in the Windows update process ( CVE-2024-21302 and CVE-2024-38202 ) that could be weaponized to rollback an up-to-date Windows software to an older version containing unpatched security vulnerabilities. The exploit materialized in the form of a tool dubbed Windows Downdate, which, per Leviev, could be used to hijack the Windows Update process to craft fully undetectable, persistent, and irreversible downgrades on...

A security flaw impacting the Wi-Fi Test Suite could enable unauthenticated local attackers to execute arbitrary code with elevated privileges. The CERT Coordination Center (CERT/CC) said the susceptible code from the Wi-Fi Alliance has been found deployed on Arcadyan FMIMG51AX000J routers. The vulnerability is being tracked as CVE-2024-41992. "This flaw allows an unauthenticated local attacker to exploit the Wi-Fi Test Suite by sending specially crafted packets, enabling the execution of arbitrary commands with root privileges on the affected routers," the CERT/CC said in an advisory released Wednesday. Wi-Fi Test Suite is an integrated platform developed by the Wi-Fi Alliance that automates testing Wi-Fi components or devices. While open-source components of the toolkit are publicly available , the full package is available only to its members.  SSD Secure Disclosure, which released details of the flaw back in August 2024, described it as a case of command injectio...

Cybersecurity researchers have disclosed a security flaw impacting Amazon Web Services (AWS) Cloud Development Kit (CDK) that could have resulted in an account takeover under specific circumstances. "The impact of this issue could, in certain scenarios, allow an attacker to gain administrative access to a target AWS account, resulting in a full account takeover," Aqua researchers Ofek Itach and Yakir Kadkoda said in a report shared with The Hacker News. Following responsible disclosure on June 27, 2024, the issue was addressed by the project maintainers in CDK version 2.149.0 released in July. AWS CDK is an open-source software development framework for defining cloud application resources using Python, TypeScript, or JavaScript and provisioning them via CloudFormation. The problem identified by Aqua builds upon prior findings from the cloud security firm about shadow resources in AWS, and how predefined naming conventions for AWS Simple Storage Service (S3) buckets ...

Cisco on Wednesday said it has released updates to address an actively exploited security flaw in its Adaptive Security Appliance (ASA) that could lead to a denial-of-service (DoS) condition. The vulnerability, tracked as CVE-2024-20481 (CVSS score: 5.8), affects the Remote Access VPN (RAVPN) service of Cisco ASA and Cisco Firepower Threat Defense (FTD) Software. Arising due to resource exhaustion, the security flaw could be exploited by unauthenticated, remote attackers to cause a DoS of the RAVPN service. "An attacker could exploit this vulnerability by sending a large number of VPN authentication requests to an affected device," Cisco said in an advisory. "A successful exploit could allow the attacker to exhaust resources, resulting in a DoS of the RAVPN service on the affected device." Restoration of the RAVPN service may require a reload of the device depending on the impact of the attack, the networking equipment company added. While there are no dire...

The North Korean threat actor known as Lazarus Group has been attributed to the zero-day exploitation of a now-patched security flaw in Google Chrome to seize control of infected devices. Cybersecurity vendor Kaspersky said it made the discovery after it came across a novel attack chain in May 2024 that targeted the personal computer of an unnamed Russian national with the Manuscrypt backdoor. This entails triggering the zero-day exploit simply upon visiting a fake game website ("detankzone[.]com") that was aimed at individuals in the cryptocurrency sector. The campaign is estimated to have commenced in February 2024. "On the surface, this website resembled a professionally designed product page for a decentralized finance (DeFi) NFT-based (non-fungible token) multiplayer online battle arena (MOBA) tank game, inviting users to download a trial version," Kaspersky researchers Boris Larin and Vasily Berdnikov said . "But that was just a disguise. Under the...