Rose debug info
---------------

how human behavior affects security

Later Ctrl + ↑

Programmer’s Digest #110

11/20/2024-11/27/2024 Critical “Array Networks” Flaw, PyPI Attack, Palo Alto Networks Firewalls Compromised And More.

1. CISA Urges Agencies to Patch Critical “Array Networks” Flaw Amid Active

AttacksCISA has added a critical vulnerability in Array Networks AG and vxAG secure access gateways (CVE-2023-28461) to its Known Exploited Vulnerabilities (KEV) catalog after reports of active exploitation. This flaw (CVSS score: 9.8) allows remote code execution without authentication via a vulnerable URL and was patched in March 2023 (version 9.4.0.484). Trend Micro linked the vulnerability to China-based cyber espionage group Earth Kasha, which targets Japanese entities and, increasingly, Taiwan, India, and Europe. Earth Kasha has exploited flaws in Array AG, Proself, and Fortinet FortiOS for initial access. ESET recently exposed a campaign by the group using the upcoming World Expo 2025 as a lure to deliver malware. CISA advises agencies to apply patches by December 16, 2024. Over 440,000 internet-exposed systems remain at risk.

2. PyPI Attack: ChatGPT, Claude Impersonators Deliver JarkaStealer via Python Libraries

Researchers have identified two malicious packages on the Python Package Index (PyPI), gptplus and claudeai-eng, impersonating popular AI models OpenAI ChatGPT and Anthropic Claude to distribute the information stealer JarkaStealer. Uploaded by “Xeroline” in November 2023, the packages attracted over 3,500 downloads before being removed. Marketed as tools to access GPT-4 Turbo and Claude AI APIs, they concealed malicious code that deployed malware upon installation. The packages’ __init__.py file contained encoded code to download a Java-based stealer, JavaUpdater.jar, from GitHub and install Java Runtime Environment if needed. Once executed, JarkaStealer harvested sensitive data, including browser data, screenshots, and app session tokens (Telegram, Discord, Steam). The stolen data was archived, sent to the attacker’s server, and deleted from the victim’s system. Sold as malware-as-a-service (MaaS) for $20–$50 on Telegram, JarkaStealer’s source code has also leaked online. 

3. 2,000 Palo Alto Networks Firewalls Compromised

A recent campaign exploiting two vulnerabilities has compromised around 2,000 Palo Alto Networks firewalls, according to Shadowserver researchers. The flaws include a critical authentication bypass (CVE-2024-0012, severity 9.3) and a medium-severity privilege escalation bug (CVE-2024-9474, severity 6.9), which can be chained for attacks. CVE-2024-0012 allows unauthenticated attackers with access to the management interface to gain admin privileges, tamper with configurations, or exploit CVE-2024-9474. The flaws, disclosed earlier in November, affect certain PAN-OS 10.2–11.2 deployments on PA-Series, VM-Series, CN-Series, and Panorama devices but not Cloud NGFW or Prisma Access.

Palo Alto Networks disputes Shadowserver’s numbers, stating fewer than 0.5% of customer firewalls have internet-exposed interfaces and that the impact is “limited.” The company emphasizes securing management interfaces to reduce risks. 

4. Decades-Old Security Vulnerabilities Found in Ubuntu’s Needrestart Package

The Qualys Threat Research Unit (TRU) has uncovered five Local Privilege Escalation (LPE) vulnerabilities in Ubuntu’s needrestart package, enabling local attackers to gain root privileges without user interaction. The flaws, introduced in version 0.8 (April 2014), affect Ubuntu Server systems with needrestart installed by default since version 21.04.

The vulnerabilities are tracked as CVE-2024-48990, CVE-2024-48991, CVE-2024-48992, CVE-2024-10224, and CVE-2024-11003. They allow attackers to execute arbitrary code as root by exploiting issues with interpreter environment variables (Python/Ruby) or race conditions.

Qualys warns that these vulnerabilities, with CVSS scores up to 7.8, are highly exploitable and could soon see public exploits, posing severe risks like unauthorized access, malware, and system compromise.

Mitigation includes disabling interpreter scanning by adding $nrconf{interpscan} = 0; to /etc/needrestart/needrestart.conf. Enterprises should update needrestart immediately to avoid operational disruptions and data breaches.

4 mo   digest   programmers'

Programmer’s Digest #109

11/13/2024-11/20/2024 High-Severity Flaw in PostgreSQL, Oracle Agile PLM Zero-Day Vulnerability, Critical WordPress Plugin Vulnerability And More.

1. Oracle Agile PLM Zero-Day Vulnerability Exploited In The Wild

Oracle has issued a security alert for a critical vulnerability (CVE-2024-21287) in its Agile Product Lifecycle Management (PLM) Framework, currently being actively exploited. The flaw, with a CVSS score of 7.5, affects version 9.3.6 and allows unauthenticated attackers to remotely access and download sensitive files via HTTP or HTTPS. Exploiting this vulnerability could grant attackers unauthorized access to critical data under the PLM application’s privileges.Oracle confirmed active exploitation and has released a security patch. Customers are urged to apply updates immediately and monitor for unauthorized activity.Organizations should act promptly to secure systems against this high-severity zero-day vulnerability.

2. Urgent: Critical WordPress Plugin Vulnerability Exposes Over 4 Million Sites

A critical vulnerability (CVE-2024-10924, CVSS 9.8) in the Really Simple Security plugin for WordPress could allow attackers to gain full administrative access to affected sites. The flaw impacts versions 9.0.0 to 9.1.1.1 of the plugin, used by over 4 million websites. It stems from improper handling of user authentication in the “check_login_and_get_user” function, enabling unauthenticated attackers to bypass two-factor authentication. The vulnerability, disclosed by Wordfence on November 6, 2024, has been patched in version 9.1.2. To mitigate risks, WordPress collaborated with the plugin developers to force-update all affected sites. Separately, another flaw (CVE-2024-10470, CVSS 9.8) in the WPLMS Learning Management System plugin allows unauthenticated users to read or delete files, potentially enabling site takeovers.These incidents highlight the importance of immediate patching and maintaining updated WordPress plugins to protect against severe exploitation.

3. Palo Alto Networks Confirms Zero-Day Exploitation in PAN-OS Firewalls

Palo Alto Networks has confirmed active exploitation of a zero-day vulnerability in its PAN-OS firewall management interface, initially reported as a potential remote code execution flaw (CVSS 9.3). The zero-day is being exploited to deploy web shells for persistent remote access. A CVE is pending assignment. Threat actors target exposed management interfaces, emphasizing the need to restrict access to trusted internal IPs. The company recommends isolating the management interface on a VLAN, using jump servers, limiting inbound IPs, and allowing only secure protocols like SSH and HTTPS. Indicators of compromise (IoCs) include malicious activity from IPs such as `136.144.17[.]*` and a specific web shell checksum. Restricting interface access significantly reduces risk, dropping the CVSS score to 7.5. Palo Alto urges immediate application of these best practices. 

Additionally, CISA added two related vulnerabilities (CVE-2024-9463 and CVE-2024-9465) to its Known Exploited Vulnerabilities catalog.

4. Researchers Warn of Privilege Escalation Risks in Google’s Vertex AI ML Platform

Researchers from Palo Alto Networks Unit 42 uncovered two critical flaws in Google’s Vertex AI platform that could enable attackers to escalate privileges and exfiltrate machine learning (ML) models. The first vulnerability exploits Vertex AI Pipelines, a feature for automating ML workflows. By manipulating custom job permissions, attackers can escalate privileges, gain unauthorized access to restricted resources, and deploy a reverse shell for backdoor access. The second flaw involves deploying a poisoned model that abuses permissions to move laterally into Kubernetes clusters. This allows attackers to exfiltrate proprietary ML models, including fine-tuned large language models (LLMs). These vulnerabilities pose serious risks, as a single malicious model could compromise an entire AI environment. Google has since patched the issues.  

5. High-Severity Flaw in PostgreSQL Allows Hackers to Exploit Environment Variables

Researchers have disclosed a critical vulnerability (CVE-2024-10979) in PostgreSQL, rated CVSS 8.8, that allows unprivileged users to alter environment variables, potentially enabling code execution or information disclosure. The flaw affects PostgreSQL’s PL/Perl extension, where improper control of environment variables (e. g., PATH) can let attackers execute arbitrary code, even without access to the server’s operating system. This could lead to severe security risks, including malicious code execution or data extraction. The issue has been addressed in PostgreSQL versions 17.1, 16.5, 15.9, 14.14, 13.17, and 12.21. Users are advised to apply the patch and restrict extension permissions, following the principle of least privilege, to minimize risk. The vulnerability was discovered by Varonis researchers Tal Peleg and Coby Abrams. More details are being withheld to allow time for users to secure their systems.

6. CISA Adds Palo Alto Networks Expedition Vulnerabilities to Exploited Catalog 

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has added multiple vulnerabilities in Palo Alto Networks’ Expedition tool to its Known Exploited Vulnerabilities (KEV) catalog. These flaws, affecting versions prior to 1.2.96, could allow attackers to access sensitive data, execute commands, and compromise firewall configurations.  

Key vulnerabilities include:  

  • CVE-2024-9463 (CVSS 9.9): Unauthenticated command injection granting root access to sensitive data.  
  • CVE-2024-9465 (CVSS 9.2):SQL injection enabling unauthorized database access and file manipulation.  
  • CVE-2024-9464 (CVSS 9.3): Authenticated OS command injection exposing credentials and API keys.  

Researchers from Horizon3 shared proof-of-concept exploits and Indicators of Compromise (IOCs). Palo Alto recommends restricting Expedition’s access to trusted users and checking for compromise.  Federal agencies must address these flaws by December 5, 2024, per CISA’s Binding Operational Directive, and private organizations are advised to follow suit to protect their networks.

5 mo   digest   programmers'

Programmer’s Digest #108

11/06/2024-11/13/2024 Patched Critical Vulnerability in Industrial Networking Solution, Security Flaws in Popular ML Toolkits, Potential PAN-OS RCE Vulnerability And More.

1. Cisco Patches Critical Vulnerability in Industrial Networking Solution

Cisco recently issued patches for multiple vulnerabilities across its enterprise products, including a critical bug in Unified Industrial Wireless software. The critical flaw, tracked as CVE-2024-20418 with a maximum CVSS score of 10, lets remote attackers execute commands with root privileges. This vulnerability exists due to improper input validation in the software’s web management interface, enabling crafted HTTP requests to bypass security controls. Devices affected include Catalyst IW9165D, IW9165E, and IW9167E access points with Ultra-Reliable Wireless Backhaul mode enabled. Cisco urges users to upgrade to version 17.15.1 to mitigate the risk.

Cisco also patched CVE-2024-20536, a high-severity vulnerability in the Nexus Dashboard Fabric Controller that could allow SQL command execution, and CVE-2024-20484 in Enterprise Chat and Email, which could cause denial-of-service. No known exploits of these flaws have been reported, and further details are available in Cisco’s security advisories.

2. Microsoft Confirms Zero-Day Exploitation of Task Scheduler Flaw

Microsoft’s security team released fixes for over 90 Windows vulnerabilities, highlighting two zero-day flaws already exploited in attacks. One flaw, CVE-2024-49039, is a privilege escalation issue in Windows Task Scheduler that could allow code execution from low-privilege applications. Discovered by Google’s Threat Analysis Group, it has a severity score of 8.8 and is suspected to have been used in targeted attacks. Another critical flaw, CVE-2024-43451, exposes a user’s NTLMv2 hash, enabling attackers to impersonate the user. Minimal interaction, like a single-click on a malicious file, could trigger this vulnerability.

The patch rollout also addresses a critical Windows Kerberos vulnerability (CVE-2024-43639) and a .NET/Visual Studio bug (CVE-2024-43498), both carrying severity scores of 9.8 and risking remote code execution. Additionally, Adobe issued fixes for 48 bugs across various products, including urgent patches for Adobe Commerce and other platforms vulnerable to code execution.

3. Security Flaws in Popular ML Toolkits Enable Server Hijacks, Privilege Escalation

Cybersecurity researchers have identified nearly two dozen vulnerabilities across 15 open-source machine learning (ML) projects, exposing critical security weaknesses. These flaws, which include issues on both server- and client-sides that could allow attackers to hijack ML model registries, databases, and pipelines. Key vulnerabilities include CVE-2024-7340, a directory traversal flaw in Weave ML that allows privilege escalation, and CVE-2024-6507, a command injection vulnerability in Deep Lake’s database that enables malicious command execution. Other severe flaws affect ZenML, Vanna.AI, and Mage AI, allowing unauthorized privilege elevation and code execution. These flaws highlight the risk of attacks such as ML model backdooring and data poisoning, which could lead to significant breaches.

4. Palo Alto Networks Warns of Potential PAN-OS RCE Vulnerability

Palo Alto Networks recently warned customers about a possible remote code execution vulnerability in the PAN-OS management interface of its next-generation firewalls. Although the company has not yet confirmed details of the vulnerability or detected active exploitation, it advises customers to restrict access to the interface by allowing only trusted internal IPs. For enhanced security, Palo Alto suggests isolating the management interface on a dedicated VLAN, using jump servers, and limiting IP access to approved devices. Cortex Xpanse and Cortex XSIAM customers can monitor for potential exposure via the Firewall Admin Login attack surface rule. Meanwhile, CISA has warned of ongoing attacks on a critical flaw in Palo Alto’s Expedition tool (CVE-2024-5910), which allows attackers to reset admin credentials. Exploits for this and related vulnerabilities, including CVE-2024-9464, have been demonstrated, potentially allowing unauthorized command execution.

5. New Flaws in Citrix Virtual Apps Enable RCE Attacks via MSMQ Misconfiguration

Cybersecurity researchers have found security flaws in Citrix Virtual Apps and Desktops that could allow remote code execution (RCE) through the Session Recording component, which logs user activity for compliance purposes. The vulnerability arises from a misconfigured Microsoft Message Queuing (MSMQ) instance, allowing unauthorized access via BinaryFormatter deserialization. Notably, Citrix stated that successful exploitation requires an authenticated user in the same domain, and patches are available in newer versions, including updates for LTSR releases. The flaws are tracked as CVE-2024-8068 and CVE-2024-8069. Microsoft has discouraged using BinaryFormatter due to security risks, removing it in .NET 9 for safety. Researchers warn the vulnerability could lead to “point-click-full-takeover” attacks. Shadowserver Foundation has observed potential exploitation attempts and strongly advises immediate patching to secure affected systems.

5 mo   digest   programmers'
Earlier Ctrl + ↓