12/28/2022-01/04/2023. PyTorch Machine Learning Framework Compromised, WordPress Security Alert, Exploitation of JasperReports Vulnerabilities And More

1. PyTorch Machine Learning Framework Compromised with Malicious Dependency

The maintainers of the PyTorch package have warned users who have installed the nightly builds of the library between December 25, 2022, and December 30, 2022, to uninstall and download the latest versions following a dependency confusion attack. The PyTorch team said that it became aware of the malicious dependency on December 30. The supply chain attack entailed uploading the malware-laced copy of a legitimate dependency named torchtriton to the Python Package Index (PyPI) code repository. Since package managers like pip check public code registries such as PyPI for a package before private registries, it allowed the fraudulent module to be installed on users’ systems as opposed to the actual version pulled from the third-party index. As mitigations, torchtriton has been removed as a dependency and replaced with pytorch-triton. A dummy package has also been registered on PyPI as a placeholder to prevent further abuse.

2. WordPress Security Alert: New Linux Malware Exploiting Over Two Dozen CMS Flaws

WordPress sites are being targeted by a previously unknown strain of Linux malware that exploits flaws in over two dozen plugins and themes to compromise vulnerable systems. If sites use outdated versions of such add-ons, lacking crucial fixes, the targeted web pages are injected with malicious JavaScripts. As a result, when users click on any area of an attacked page, they are redirected to other sites. The attacks involve weaponizing a list of known security vulnerabilities in 19 different plugins and themes that are likely installed on a WordPress site, using it to deploy an implant that can target a specific website to further expand the network. It’s also capable of injecting JavaScript code retrieved from a remote server in order to redirect the site visitors to an arbitrary website of the attacker’s choice. WordPress users are recommended to keep all the components of the platform up-to-date, including third-party add-ons and themes. It’s also advised to use strong and unique logins and passwords to secure their accounts.
 

3. CISA Warns of Active Exploitation of JasperReports Vulnerabilities

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has added two years-old security flaws impacting TIBCO Software’s JasperReports product to its Known Exploited Vulnerabilities (KEV) catalog, citing evidence of active exploitation. The flaws, tracked as CVE-2018-5430 (CVSS score: 7.7) and CVE-2018-18809 (CVSS score: 9.9), were addressed by TIBCO in April 2018 and March 2019, respectively. The first of the two issues, CVE-2018-5430, relates to an information disclosure bug in the server component that could enable an authenticated user to gain read-only access to arbitrary files, including key configurations.

JasperReports Vulnerabilities

The impact includes the possible read-only access by authenticated users to web application configuration files that contain the credentials used by the server. Those credentials could then be used to affect external systems accessed by the JasperReports Server.
CVE-2018-18809 is a directory traversal vulnerability in the JasperReports Library that could permit web server users to access sensitive files on the host, potentially making it possible for an attacker to steal credentials and break into other systems. 
 

4. Thousands of Citrix Servers Still Unpatched for Critical Vulnerabilities

Thousands of Citrix Application Delivery Controller (ADC) and Gateway endpoints remain vulnerable to two critical security flaws disclosed by the company over the last few months. The issues in question are CVE-2022-27510 and CVE-2022-27518 (CVSS scores: 9.8). While CVE-2022-27510 relates to an authentication bypass that could be exploited to gain unauthorized access to Gateway user capabilities, CVE-2022-27518 concerns a remote code execution bug that could enable the takeover of affected systems. Citrix  warned that CVE-2022-27518 is being actively exploited in the wild by threat actors, including the China-linked APT5 state-sponsored group.  According to a new analysis from NCC Group’s Fox-IT research team, thousands of internet-facing Citrix servers are still unpatched, making them an attractive target for hacking crews. This includes over 3,500 Citrix ADC and Gateway servers running version 12.1-65.21 that are susceptible to CVE-2022-27518, as well as more than 500 servers running 12.1-63.22 that are vulnerable to both flaws. A majority of the servers, amounting to no less than 5,000, are running 13.0-88.14, a version that’s immune to CVE-2022-27510 and CVE-2022-27518.

5. Several DoS, Code Execution Vulnerabilities Found in Rockwell Automation Controllers

 Organizations using controllers made by Rockwell Automation have been informed recently about several potentially serious vulnerabilities. One flaw is CVE-2022-3156, which impacts the Studio 5000 Logix Emulate controller emulation software. The vulnerability is caused by a misconfiguration that results in users being granted elevated permissions on certain product services. An attacker could exploit the weakness for remote code execution.
The second vulnerability is CVE-2022-3157, which affects CompactLogix, GuardLogix (including Compact), and ControlLogix controllers. An attacker can exploit the flaw to launch a denial-of-service (DoS) attack against a device by sending specially crafted CIP requests that cause a “major non-recoverable fault”. The remaining vulnerabilities impact MicroLogix 1100 and 1400 programmable logic controllers (PLCs). One of the security holes, CVE-2022-46670, is a stored cross-site scripting (XSS) issue in the embedded webserver that can be exploited for remote code execution without authentication. The second bug, CVE-2022-3166, is a clickjacking issue that can be exploited by an attacker with network access to the affected device to cause a DoS condition for the webserver application. 
The first two vulnerabilities have been patched with updates. For the last two issues, the vendor has made available mitigations that should prevent attacks.

6. Critical Vulnerabilities Patched in Synology Routers

Taiwan-based networking and storage solutions provider Synology has informed customers about the availability of patches for several critical vulnerabilities, including flaws likely exploited recently at the Pwn2Own hacking contest. The company published two new critical advisories in late December. One of them describes an internally discovered vulnerability affecting Synology VPN Plus Server, which turns routers into an advanced VPN server.
The security hole, tracked as CVE-2022-43931, is an out-of-bounds write issue in the remote desktop functionality of VPN Plus Server. It can allow a remote attacker to execute arbitrary commands. The second advisory describes multiple vulnerabilities impacting the Synology Router Manager (SRM), the operating system that powers the firm’s routers. The flaws can be exploited for arbitrary command execution, denial-of-service (DoS) attacks, and reading arbitrary files. 

22-28/12/2022. Hackers Breach Okta’s GitHub Repositories, W4SP Stealer Discovered in Multiple PyPI Packages, Security Flaws in Ghost CMS Blogging Software And More

1. W4SP Stealer Discovered in Multiple PyPI Packages Under Various Names

Threat actors have published another round of malicious packages to Python Package Index (PyPI) with the goal of delivering information-stealing malware on compromised developer machines. Interestingly, while the malware goes by a variety of names like ANGEL Stealer, Celestial Stealer, Fade Stealer, Leaf $tealer, PURE Stealer, Satan Stealer, and @skid Stealer, cybersecurity company Phylum found them all to be copies of W4SP Stealer. W4SP Stealer primarily functions to siphon user data, including credentials, cryptocurrency wallets, Discord tokens, and other files of interest. For some reason, each deployment appears to have simply tried to do a find/replace of the W4SP references in exchange for some other seemingly arbitrary name. The campaign distributing W4SP Stealer gained traction around October 2022.  Since then dozens of additional bogus packages containing W4SP Stealer have been published on PyPI by the persistent threat actors. It’s worth noting that previous versions of the attack chains have been spotted fetching next-stage Python code directly from a public GitHub repository that then drops the credential stealer. 

2.  LastPass Admits to Severe Data Breach, Encrypted Password Vaults Stolen

The August 2022 security breach of LastPass may have been more severe than previously disclosed by the company. The popular password management service  revealed that malicious actors obtained a trove of personal information belonging to its customers that include their encrypted password vaults by using data siphoned from the earlier break-in. Also stolen is “basic customer account information and related metadata including company names, end-user names, billing addresses, email addresses, telephone numbers, and the IP addresses from which customers were accessing the LastPass service”.  The August 2022 incident involved the miscreants accessing source code and proprietary technical information from its development environment via a single compromised employee account. LastPass said this permitted the unidentified attacker to obtain credentials and keys that were subsequently leveraged to extract information from a backup stored in a cloud-based storage service, which it emphasized is physically separate from its production environment.  LastPass confirmed that the security lapse did not involve access to unencrypted credit card data, as this information was not archived in the cloud storage container. The company did not divulge how recent the backup was, but warned that the threat actor “may attempt to use brute-force to guess your master password and decrypt the copies of vault data they took,” as well as target customers with social engineering and credential stuffing attacks.

3.  GuLoader Malware Utilizing New Techniques to Evade Security Software

Cybersecurity researchers have exposed a wide variety of techniques adopted by an advanced malware downloader called GuLoader to evade security software. New shellcode anti-analysis technique attempts to thwart researchers and hostile environments by scanning entire process memory for any virtual machine (VM)-related strings. GuLoader, also called CloudEyE, is a Visual Basic Script (VBS) downloader that’s used to distribute remote access trojans such as Remcos on infected machines. A recent GuLoader sample unearthed by CrowdStrike exhibits a three-stage process wherein the VBScript is designed to deliver a next-stage that performs anti-analysis checks before injecting shellcode embedded within the VBScript into memory. The shellcode, besides incorporating the same anti-analysis methods, downloads a final payload of the attacker’s choice from a remote server and executes it on the compromised host. The shellcode employs several anti-analysis and anti-debugging tricks at every step of execution, throwing an error message if the shellcode detects any known analysis of debugging mechanisms. This includes anti-debugging and anti-disassembling checks to detect the presence of a remote debugger and breakpoints, and if found, terminate the shellcode. The shellcode also features scans for virtualization software. An added capability is what the cybersecurity company calls a “redundant code injection mechanism” to avoid NTDLL.dll hooks implemented by endpoint detection and response (EDR) solutions. In a nutshell, the method involves using assembly instructions to invoke the necessary windows API function to allocate memory (i.e., NtAllocateVirtualMemory) and inject arbitrary shellcode into that location via process hollowing.

4.  Two New Security Flaws Reported in Ghost CMS Blogging Software

Cybersecurity researchers have detailed two security flaws in the JavaScript-based blogging platform known as Ghost, one of which could be abused to elevate privileges via specially crafted HTTP requests. Tracked as CVE-2022-41654 (CVSS score: 9.6), the authentication bypass vulnerability allows unprivileged users (i.e., members) to make unauthorized modifications to newsletter settings. Cisco Talos, which discovered the shortcoming, said it could enable a member to change the system-wide default newsletter that all users are subscribed to by default. Even worse, the ability of a site administrator to inject JavaScript into the newsletter by default could be exploited to trigger the creation of arbitrary administrator accounts when attempting to edit the newsletter.  The CMS platform blamed the bug due to a “gap” in its API validation, adding it found no evidence that the issue has been exploited in the wild. Also patched by Ghost is an enumeration vulnerability in the login functionality (CVE-2022-41697, CVSS score: 5.3) that could lead to the disclosure of sensitive information.

5. Zerobot Malware Now Spreads By Exploiting Apache Vulnerabilities

The Zerobot botnet has been upgraded to infect new devices by exploiting security vulnerabilities affecting Internet-exposed and unpatched Apache servers. The Microsoft Defender for IoT research team also observed that this latest version adds new distributed denial-of-service (DDoS) capabilities. Zerobot has been under active development since at least November, with new versions adding new modules and features to expand the botnet’s attack vectors and make it easier to infect new devices, including firewalls, routers, and cameras. Since early December, the malware’s developers have removed modules that targeted phpMyAdmin servers, Dasan GPON home routers, and D-Link DSL-2750B wireless routers with year-old exploits. The update spotted by Microsoft adds newer exploits to the malware’s toolkit, enabling it to target seven new types of devices and software, including unpatched Apache and Apache Spark servers. 
The  list of modules added to Zerobot 1.1 includes:

• CVE-2017-17105: Zivif PR115-204-P-RS
• CVE-2019-10655: Grandstream
• CVE-2020-25223: WebAdmin of Sophos SG UTM
• CVE-2021-42013: Apache
• CVE-2022-31137: Roxy-WI

Microsoft researchers have also found new evidence that Zerobot propagates by compromising devices with known vulnerabilities that are not included in the malware binary, such as CVE-2022-30023, a command injection vulnerability in Tenda GPON AC1200 routers.

6. Critical Vulnerability in Premium Gift Cards WordPress Plugin Exploited in Attacks

 Defiant’s Wordfence team warns of a critical-severity vulnerability in the YITH WooCommerce Gift Cards premium WordPress plugin being exploited in attacks. The YITH WooCommerce Gift Cards plugin allows online merchants to create gift cards that their customers can purchase for their friends to use on the ecommerce store. The premium plugin has more than 50,000 installations, its developer says. Tracked as CVE-2022-45359 (CVSS score of 9.8), the exploited vulnerability was reported in November and a patch for it was made available soon after. The issue is described as an arbitrary file upload, allowing attackers to upload executable files to the WordPress sites that use a vulnerable version of the plugin. No authentication is required for successful exploitation. According to the WordPress security firm, an attacker can exploit the vulnerability to place a backdoor on a vulnerable installation, gain remote code execution (RCE), and potentially take over the site. Site admins can identify signs of an attack by checking their logs for POST requests to wp-admin/admin-post.php.
According to Wordfence, observed attacks came from hundreds of IP addresses, but only two IPs were responsible for the majority of exploitation attempts. Site admins are advised to update to YITH WooCommerce Gift Cards premium version 3.20.0 or newer, which contain patches for this vulnerability.

15-21/12/2022 Malicious ‘SentinelOne’ PyPI package, Hackers bombard PyPi platform, Veeam Backup and Replication Vulnerabilities, And More

1.  Hackers Bombard Open Source Repositories with Over 144,000 Malicious Packages

NuGet, PyPi, and npm ecosystems are the target of a new campaign that has resulted in over 144,000 packages being published by unknown threat actors.The packages were part of a new attack vector, with attackers spamming the open source ecosystem with packages containing links to phishing campaigns. Of the 144,294 phishing-related packages that were detected, 136,258 were published on NuGet, 7,824 on PyPi, and 212 on npm. The offending libraries have since been unlisted or taken down. The fake packages themselves claimed to provide hacks, cheats, and free resources in an attempt to trick users into downloading them. The URLs to the rogue phishing pages were embedded in the package description. In all, the massive campaign encompassed more than 65,000 unique URLs on 90 domains.

2. CISA Alert: Veeam Backup and Replication Vulnerabilities Being Exploited in Attacks

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has added two vulnerabilities impacting Veeam Backup & Replication software to its Known Exploited Vulnerabilities (KEV) Catalog, citing evidence of active exploitation in the wild. The now-patched critical flaws, tracked as CVE-2022-26500 and CVE-2022-26501, are both rated 9.8 on the CVSS scoring system, and could be leveraged to gain control of a target system. The Veeam Distribution Service (TCP 9380 by default) allows unauthenticated users to access internal API functions. A remote attacker may send input to the internal API which may lead to uploading and executing of malicious code. Both the issues that impact product versions 9.5, 10, and 11 have been addressed in versions 10a and 11a. Users of Veeam Backup & Replication 9.5 are advised to upgrade to a supported version.

3. Researchers Discover Malicious PyPI Package Posing as SentinelOne SDK to Steal Data

Cybersecurity researchers have discovered a new malicious package on the Python Package Index (PyPI) repository that impersonates a software development kit (SDK) for SentinelOne, a major cybersecurity company, as part of a campaign dubbed SentinelSneak. The package, named SentinelOne and now taken down, is said to have been published between December 8 and 11, 2022, with nearly two dozen versions pushed in quick succession over a period of two days. It claims to offer an easier method to access the company’s APIs, but harbors a malicious backdoor that’s engineered to amass sensitive information from development systems, including access credentials, SSH keys, and configuration data. What’s notable about the fraudulent package is it mimics a legitimate SDK that’s offered by SentinelOne to its customers, potentially tricking developers into downloading the module from PyPI. It’s not immediately clear if the package was weaponized as part of an active supply chain attack, although it has been downloaded more than 1,000 times prior to its removal.

4. Glupteba Botnet Continues to Thrive Despite Google’s Attempts to Disrupt It

 The operators of the Glupteba botnet resurfaced in June 2022 as part of a renewed and “upscaled” campaign, months after Google disrupted the malicious activity. The ongoing attack is suggestive of the malware’s resilience in the face of takedowns. In addition, there was a tenfold increase in TOR hidden services being used as C2 servers since the 2021 campaign. The malware, which is distributed through fraudulent ads or software cracks, is also equipped to retrieve additional payloads that enable it to steal credentials, mine cryptocurrencies, and expand its reach by exploiting vulnerabilities in IoT devices from MikroTik and Netgear. Specifically, the botnet is designed to search the public Bitcoin blockchain for transactions related to wallet addresses owned by the threat actor so as to fetch the encrypted C2 server address. This is made possible by the OP_RETURN opcode that enables storage of up to 80 bytes of arbitrary data within the signature script.

5. New Agenda Ransomware Variant, Written in Rust, Aiming at Critical Infrastructure

A Rust variant of a ransomware strain known as Agenda has been observed in the wild, making it the latest malware to adopt the cross-platform programming language after BlackCat, Hive, Luna, and RansomExx. Agenda is a ransomware-as-a-service (RaaS) group that has been linked to a spate of attacks primarily targeting manufacturing and IT industries. It expands on the idea of partial encryption (aka intermittent encryption) by configuring parameters that are used to determine the percentage of file content to be encrypted. An analysis of the ransomware binary reveals that encrypted files are given the extension “MmXReVIxLV,” before proceeding to drop the ransom note in every directory. In addition, the Rust version of Agenda is capable of terminating the Windows AppInfo process and disabling User Account Control (UAC), the latter of which helps mitigate the impact of malware by requiring administrative access to launch a program or task. At present, its threat actors appear to be migrating their ransomware code to Rust as recent samples still lack some features seen in the original binaries written in the Golang variant of the ransomware.

6. Samba Issues Security Updates to Patch Multiple High-Severity Vulnerabilities

Samba has released software updates to remediate multiple vulnerabilities that, if successfully exploited, could allow an attacker to take control of affected systems. The high-severity flaws, tracked as CVE-2022-38023, CVE-2022-37966, CVE-2022-37967, and CVE-2022-45141, have been patched in versions 4.17.4, 4.16.8 and 4.15.13 released on December 15, 2022. Samba is an open source Windows interoperability suite for Linux, Unix, and macOS operating systems that offers file server, printing, and Active Directory services.
A brief description of each of the weaknesses is below:

• CVE-2022-38023 (CVSS score: 8.1) – Use of weak RC4-HMAC Kerberos encryption type in the Netlogon Secure Channel
• CVE-2022-37966 (CVSS score: 8.1) – An elevation of privilege vulnerability in Windows Kerberos RC4-HMAC
  An unauthenticated attacker could conduct an attack that could leverage cryptographic protocol vulnerabilities in RFC 4757 (Kerberos encryption type RC4-HMAC-MD5) and MS-PAC (Privilege Attribute Certificate Data Str

7. Okta Says Its GitHub Account Hacked, Source Code Stolen

Okta, a leading provider of authentication services and Identity and Access Management (IAM) solutions, says that its private GitHub source code repositories were hacked this month. 
Earlier this month, GitHub alerted Okta of suspicious access to Okta’s code repositories, states the notification. Despite stealing Okta’s source code, attackers did not gain unauthorized access to the Okta service or customer data, says the company. Okta’s “HIPAA, FedRAMP or DoD customers” remain unaffected as the company “does not rely on the confidentiality of its source code as a means to secure its services.” As such, no customer action is needed.

8. GitHub Announces Free Secret Scanning for All Public Repositories

GitHub said it is making available its secret scanning service to all public repositories on the code hosting platform for free. Secret scanning alerts notify you directly about leaked secrets in your code. Secret scanning is designed to examine repositories for access tokens, private keys, credentials, API keys, and other secrets in over 200 formats that may have been accidentally committed, and generate alerts to prevent their misuse. The security option was previously limited to repositories owned by organizations that use GitHub Enterprise Cloud and have a GitHub Advanced Security license. For customers of GitHub Advanced Security, the protections go a step further by performing the scans for exposed secrets, including custom patterns, during code pushes. The Microsoft subsidiary also said it’s planning to turn on two-factor authentication requirements for “distinct groups of users” starting March 2023 with the goal of expanding it to all GitHub users by the end of next year.