Enterprise-grade Juniper Networks routers have become the target of a custom backdoor as part of a campaign dubbed J-magic.
According to the Black Lotus Labs team at Lumen Technologies, the activity is so named for the fact that the backdoor continuously monitors for a "magic packet" sent by the threat actor in TCP traffic. 
"J-magic campaign marks the rare occasion of malware designed

Custom Backdoor Exploiting Magic Packet Vulnerability in Juniper Routers

Advanced persistent threat group PlushDaemon, active since 2019, is using a sophisticated modular backdoor to collect data from infected systems in South Korea.

Source: BeeBright via Shutterstock

A newly discovered Chinese threat group has targeted a South Korean VPN developer with a supply chain attack aimed at deploying a custom backdoor to collect data for cyber-espionage purposes.

The group, dubbed PlushDaemon by the researchers at ESET Research who discovered it, typically aims to hijack legitimate updates of Chinese applications in its malicious operations "by redirecting traffic to attacker-controlled servers," according to a blog post by ESET researcher Facundo Muñoz published on Jan. 22. "Additionally, we have observed the group gaining access via vulnerabilities in legitimate web servers," he wrote.

However, the researchers also discovered the group in May 2024 planting malicious code in an NSIS installer for the Windows version of the VPN software of South Korean company IPany, representing a departure from its typical operations, they said. ESET notified IPany and the malicious installer was removed from the company's website.

PlushDaemon has been active since at least 2019, engaging in cyberespionage operations against individuals and entities in mainland China, Taiwan, Hong Kong, South Korea, the US, and New Zealand. The group is the exclusive user of several types of malware in its malicious activities, mostly notably a custom, modular backdoor for collecting various data from infected machines, called SlowStepper for Windows, according to ESET.

**Atypical Supply-Chain Attack**

The first sign of the supply-chain attack came in May 2024, when ESET researchers noticed detections of malicious code in an NSIS installer for Windows that users from South Korea had downloaded from the IPany website.

"The victims appear to have manually downloaded a ZIP archive containing a malicious NSIS installer from the URL https://ipany\[.\]kr/download/IPanyVPNsetup.zip," Muñoz wrote. However, the researchers didn't find suspicious code on the download page "to produce targeted downloads, for example by geofencing to specific targeted regions or IP ranges." This led them to believe that "anyone using the IPany VPN might have been a valid target."

Several users attempted to install the Trojanized software in the network of a semiconductor company and an unidentified software development company in South Korea. Further research found even older cases of infection via the campaign, with the two oldest coming from a victim in Japan in November 2023 and a victim in China in December 2023, the researchers said.

**SlowStepper Backdoor**

The payload in the supply chain attack is PlushDaemon's own SlowStepper backdoor, which has more than 30 modules. However, the group used a "lite" version of the backdoor in the IPany attack, which contains fewer features than other previous and newer versions, the researchers said.

The backdoor features a multistage command-and-control (C2) protocol using DNS, and is known for its ability to download and execute dozens of additional Python modules with espionage capabilities.

"Both the full and Lite versions make use of an array of tools programmed in Python and Go, which include capabilities for extensive collection of data, and spying through recording of audio and videos," Muñoz wrote.

The researchers found PlushDaemon's tools stored in a remote code repository hosted on the Chinese platform GitCode, under the LetMeGo22 account. At the time of writing, the profile was private.

**Another Chinese APT Emerges**

China already has a raft of known and active APTs that regularly and persistently engage in cyberespionage activities against the US and its allies. One of the most notable operations of late was the infiltration of US broadband provider networks by Chinese APT Salt Typhoon; however, the investigation into that incident was dealt a significant blow on Jan. 21, when President Trump, on his second day back in office, fired the cyber safety board looking into it.

However, with a new, sophisticated actor like PlushDaemon now emerging from the shadows, organizations need to be more vigilant than ever against malicious cyber activity from China, Muñoz said.

"The numerous components in the PlushDaemon toolset and its rich version history show that, while previously unknown, this China-aligned APT group has been operating diligently to develop a wide array of tools, making it a significant threat to watch for," he wrote.

To that end, ESET included a link to its GitHub repository that contains a comprehensive list of indicators of compromise (IoCs) and samples of PlushDaemon activity.

**About the Author**

Elizabeth Montalbano is a freelance writer, journalist, and therapeutic writing mentor with more than 25 years of professional experience. Her areas of expertise include technology, business, and culture. Elizabeth previously lived and worked as a full-time journalist in Phoenix, San Francisco, and New York City; she currently resides in a village on the southwest coast of Portugal. In her free time, she enjoys surfing, hiking with her dogs, traveling, playing music, yoga, and cooking.

Chinese Cyberspies Target South Korean VPN in Supply Chain Attack

Feeling creative? Have something to say about cybersecurity? Submit your caption and our panel of experts will reward the winner with a $25 gift card.

What motivates you? What will change how you do security? Send us a cybersecurity-related caption to describe the above scene and our favorite entry will win its wordsmith a $25 gift card.

Here are four convenient ways to submit your ideas before the Feb. 12 deadline:

*   Via social media: Bluesky, LinkedIn, Facebook, and X. (If you win, we'll respond to you on the same platform, requesting your email address.)
    

**Last Month's Winner**

Shout out — and a $25 gift card — to Paul Mauriks, ICT security specialist, technology solutions, at the Department of Justice and Community Safety in Victoria, Australia for sending us the winning caption for last month's "Sneaking Around" cartoon. Some noteworthy contenders included "I guarantee you there are no backdoors to our network," "Finally found a way to get rid of all that old IT gea.," and "‘Yeah, asset disposal by ‘Back Door Inc.’ I’ll do the supplier assessment thing on Monday…’" Congrats to Paul, and a big thank you to all who participated.

**About the Author**

Cartoonist

John Klossner has been drawing technology cartoons for more than 15 years. His work regularly appears in Computerworld and Federal Computer Week. His illustrations and cartoons have also been published in The New Yorker, Barron's, and The Wall Street Journal.

Name That Toon: Incentives

The US Department of Commerce will prohibit the import of components for connected vehicles from China or Russia, as the US continues to ban technology it sees as potential national security threats.

Source: Hsyn20 via Shutterstock

Smart-vehicle makers are facing supply chain disruption as the US Department of Commerce plans to enforce new regulations banning the import of connected-vehicle technology from China and Russia over cybersecurity fears.

The Commerce Department pursued new regulations after President Biden declared a national emergency over concerns that the United States had become overreliant on China for information and communications technology and services (ICTS). The rule mandates that companies and their suppliers eliminate hardware or software imported from China or Russia in their vehicle connectivity system (VCS) or in their automated driving system (ADS).

It aims to address two concerns: vulnerabilities that would allow a nation-state or criminal organization to implant a backdoor in automotive hardware or software; and the collection of data on US drivers through diagnostic features and other mechanisms, says Yoav Levy, CEO and co-founder of automotive cybersecurity provider Upstream.

"The threat is definitely real," he says. "There are many cases where cars could be hacked — including the safety elements within the cars — and there were many cases where data was stolen or leaked. ... But so far, we haven't seen something like that on a huge scale."

Related:Leveraging Behavioral Insights to Counter LLM-Enabled Hacking

The concerns come as software-defined vehicles (SDVs) shake up the automotive market, while also potentially increasing the cyberattack surface area of automobiles. In the past, vehicle makers created a variety of platforms for their different models, and the number of processors — known as electronic control units (ECUs) — quickly climbed. While the post-pandemic chip shortage slowed the shift to new platforms, manufacturers now aim to quickly reduce the number of ECUs and other hardware needed for the VCS and ADS systems. While current models, for example, can have as many as 130 ECUs, Rivian has already reduced the number of ECUs to seven in its second generation R1 vehicles.

**Wielding the Cyber-Ban Hammer**

Rivian aside, most automobiles have a wide variety of components sourced from China, raising concerns that the United States' reliance on the technologies could allow future compromises.

Banning technology from China and sanctioning Russia is nothing new, says Ivan Novikov, CEO at API security firm Wallarm. The US government has already raised cybersecurity concerns over telecommunications equipment from Huawei, Chinese-made cargo equipment at US seaports, home routers made by Chinese manufacturer TP-Link, and popular social media app TikTok.

Related:Strategic Approaches to Threat Detection, Investigation & Response

"This is kind of the next logical step," he says.

The new commerce regulations will prohibit any "transactions involving VCS hardware and covered software designed, developed, manufactured, or supplied" by people or organizations linked to China or Russia, according to a 213-page final rule, which will be put into effect after months of comments.

Yet, many implementation details remain unclear, Novikov says.

"The open question here is who will enforce the regulations, because the usual enforcement of security requirements and crash \[safety\] tests is under the Department of Transportation," he says. "It's unclear how these two agencies can work together, and how this final DoT requirements or restrictions or controls can work."

**Securing Supply Chains & the Economy?**

The impact on the supply chain will be significant, experts say. The first tier of OEMs — large US and international companies — are not the problem. Their products, however, often come from suppliers that source their own components from Chinese companies, says Alex Oyler, director for North America at industry consultancy SBD Automotive.

It's just one more way that the supply chain is currently undergoing changes, he says. Many carmakers are looking to rewrite their relationships with providers as they move to software-defined vehicles.

Related:Trusted Apps Sneak a Bug Into the UEFI Boot Process

"We're in a bit of a different phase of software-defined vehicle in the sense that OEMs are actually starting to become a lot more prescriptive in the specification of the components that they're sourcing," Oyler says. "It's more of what's called a build-to-print relationship, where they provide not the functional requirements, but requirements for the component architecture — we want this processor, we need this memory, we need this GPU."

The shift to other sources of supply will take years, with the Biden administration allowing carmakers a grace period to comply with the regulations: Software components can no longer be sourced from China and Russia starting with 2027 car models, while by 2030 car models must contain no hardware from prohibited sources.

Making such changes will not be easy, says Upstream's Levy.

"It's not that easy to replace a supplier," he says. "There are financial implications with the supply chain — maybe it's going to be more expensive, or there may be some changes to software that they would need to do for the for the new supplier — an adjustment to the architecture. ... It really depends on what they are actually going to replace."

**About the Author**

Veteran technology journalist of more than 20 years. Former research engineer. Written for more than two dozen publications, including CNET News.com, Dark Reading, MIT's Technology Review, Popular Science, and Wired News. Five awards for journalism, including Best Deadline Journalism (Online) in 2003 for coverage of the Blaster worm. Crunches numbers on various trends using Python and R. Recent reports include analyses of the shortage in cybersecurity workers and annual vulnerability trends.

US Ban on Automotive Components Could Curb Supply Chain

A highly targeted cyber-intelligence campaign adds fuel to the increasingly complex relationship between the two former Soviet states.

Source: Daniren via Alamy Stock Photo

A suspected Russia-nexus threat actor has been executing convincing spear phishing attacks against diplomatic entities in Kazakhstan.

UAC-0063, active since at least 2021, was first documented by Ukraine's Computer Emergency Response Team (CERT-UA) in 2023. With medium confidence, CERT-UA tied it to APT28 (aka Fancy Bear, Forest Blizzard, Strontium, Sofacy), from the General Staff Main Intelligence Directorate (GRU) Military Unit 26165. APT28 is best known for its high-profile attacks against Western governments: the Democratic National Committee (DNC) hack of 2016, campaigns against parliamentary bodies in Germany, Norway, and the Netherlands, and much more.

UAC-0063, specifically, has used cyber operations to collect intelligence from government entities, nongovernmental organizations (NGOs), academic institutions, and energy and defense organizations in Eastern Europe — most notably Ukraine — as well as Central Asia, including Kazakhstan, Kyrgyzstan, Tajikistan, and other countries in the vicinity, including Israel and India.

Its latest ongoing campaign, which, in a blog post, researchers from Sekoia date back to at least 2022, may fold into a broader effort by Vladimir Putin's government to gain strategic insights into, and advantage over, a former Soviet state that has sought to broaden its diplomatic horizons in recent years.

**Phishing Kazakh Diplomats**

On Oct. 16, 2024 — one month after it'd been deployed in the wild — researchers spotted a diplomatic document uploaded to VirusTotal. It appeared to be a legitimate draft of a joint declaration between the chancellor of Germany and heads of Central Asian countries.

"The first step, when you open this document, is that it asks you to enable macros," recalls Amaury Garçon, cyber threat intelligence (CTI) analyst at Sekoia Threat Detection & Research (TDR), adding that the document was obscured by "shapes" at first sight. "Some phishing documents look really ugly or have a bad shape \[at first\] — they prompt the user to enable macros, because if you don't enable macros you can't write text in the document, can't move images, etc.," he notes.

Clicking "enable" would trigger various malicious, unseen commands on a target device. While the user was made privy to the full, unadulterated lure document, in the background their security settings would be downgraded so as to remove the need for future "enable macros" prompts. Next a second, blank document was created and opened by a hidden instance of Microsoft Word. The Visual Basic (VB) code associated with this hidden document — now enabled by default, of course — dropped and executed a malicious HTML application (HTA) containing a backdoor named "HatVibe."

The purpose of HatVibe is to receive and execute code from a remote server. Though Sekoia couldn't identify the payloads associated with this phishing campaign, CERT-UA has previously observed HatVibe downloading and executing a more complex Python backdoor named "CherrySpy."

**What This Means for Kazakhstan and Russia**

Six weeks after researchers spotted the first VirusTotal upload associated with this campaign, on Nov. 27, Putin went on a two-day state visit to the country he deemed Russia's "true ally," Kazakhstan. He and Kazakhstan's president, Kassym-Jomart Tokayev, used the opportunity afforded by the Collective Security Treaty Organization (CSTO) summit to discuss various areas for economic partnership — particularly around the energy sector — and signed agreements over energy, education, and transportation.

"Central Asia is a real point of interest for Russian influence," Maxime Arquillière, senior CTI analyst at Sekoia TDR explains. "We know that Kazakhstan is a close ally, but since the beginning of the Ukraine war, Kazakhstan has distanced itself a little bit from Russia, trying to develop new connections with both Western states and also China."

Kazakhstan's centrality in the Asian continent positions it nicely as a trade bridge between China and Europe, particularly while Ukraine and Russia are consumed by war. And as Sekoia notes in its blog, the country's gradually broadening geopolitical ties are evident in recent agreements with Mongolia and Afghanistan's new Taliban government, and, most notably, its balanced position on the war in Ukraine — supporting Ukraine's right to territorial integrity without outright condemning Russia's invasion.

This latest cyber campaign, then, fits neatly into Russia's broader initiatives with regard to its Central Asian neighbor. Sekoia identified 11 lure documents in all, each one legitimate and likely having originated with Kazakhstan's Ministry of Foreign Affairs, pertaining to diplomatic business between Kazakhstan and potential partner nations.

Exactly how the threat actor obtained these documents is not known. They include, for example:

*   Letters from Kazakhstan's embassies in Afghanistan and Belgium, regarding diplomatic and economic developments.
    
*   A draft of a joint statement between Germany and Central Asian states, following a Sept. 16, 2024, summit in Astana.
    
*   Administrative reports and briefings on the Kazakh president's visits to Mongolia and New York.
    

"It's really coherent with the \[need for\] Russian intelligence to conduct this kind of cyber espionage, to know about the strategic interests between Kazakhstan and European states," Arquillière says.

**About the Author**

Nate Nelson is a writer based in New York City. He formerly worked as a reporter at Threatpost, and wrote "Malicious Life," an award-winning Top 20 tech podcast on Apple and Spotify. Outside of Dark Reading, he also co-hosts "The Industrial Security Podcast."

Russian APT Phishes Kazakh Gov't for Strategic Intel

Seven system recovery programs contained what amounted to a backdoor for injecting any untrusted file into the system startup process.

Source: Ognyan Yosifov via Alamy Stock Photo

A vulnerability in trusted system recovery programs could allow privileged attackers to inject malware directly into the system startup process in Unified Extensible Firmware Interface (UEFI) devices.

Seven real-time recovery products — Howyar SysReturn, Greenware GreenGuard, Radix SmartRecovery, Sanfong EZ-back System, WASAY eRecoveryRX, CES NeoImpact, and SignalComputer HDD King — all make use of "reloader.efi," the Microsoft-signed Extensible Firmware Interface (EFI) file at issue.

The problem, ESET explains in a new report, is that reloader.efi uses a custom loader that enables the application to load even unsigned binaries during the boot process. In essence, it's a backdoor for sneaking any kind of file into a system's startup, past UEFI Secure Boot. The issue has been assigned CVE-2024-7344, and earned a "medium" 6.5 Common Vulnerability Scoring System (CVSS) rating, as it requires administrator privileges to exploit.

**Backdoor to the UEFI Boot Process**

The standard way to load, prepare, and execute UEFI images in system memory is with the autological LoadImage and StartImage functions. The Microsoft-approved "reloader" application goes its own way, using a custom mechanism that allows it to load any binary, trusted or otherwise, at startup.

"Maybe it's a lack of secure coding awareness," Martin Smolár, malware researcher at ESET, guesses of the developers' motives in implementing the custom loader. "Or maybe it's because they found it convenient to create such a functionality. Because when a developer makes a change \[to a signed program\] they need to send it to Microsoft to get it re-signed. This means that they don't need to every time they create a new update or something like that."

Reloader.efi loads arbitrary binaries from a specific, encrypted file, "cloak.dat." When ESET decrypted cloak.dat, it found that it contained an unsigned executable primarily designed for classroom environments. "Its core function is to provide real-time system recovery, ensuring that students from different classes can work in a teacher-predefined computer environment within shared computer labs," Smolár says, though he adds that the same component might be used in other settings, like public Internet cafes. The larger point is that the unsigned executable is run during the startup process, completely bypassing UEFI Secure Boot checks.

This odd classroom recovery software is perfectly honest, but an attacker could easily swap it out for something worse. If they could just get a hold of administrator privileges on a targeted machine, an attacker could access the EFI system partition (ESP) and substitute their own malicious file in place of cloak.dat. Then all they'd need is a quick system reboot to drop any malicious file they wished into the startup process.

**Why UEFI Bugs Are So Bad**

UEFI is a kind of sacred space — a bridge between firmware and operating system, allowing a machine to boot up in the first place.

Any malware that invades this space will earn a dogged persistence through reboots, by reserving its own spot in the startup process. Security programs have a harder time detecting malware at such a low level of the system. Even more importantly, by loading first, UEFI malware will simply have a head start over those security checks that it aims to avoid. Malware authors take advantage of this order of operations by designing UEFI bootkits that can hook into security protocols, and undermine critical security mechanisms like UEFI Secure Boot or HVCI (Hypervisor-Protected Code Integrity), Windows' technology for blocking unsigned code in the kernel.

To ensure that none of this can happen, the UEFI Boot Manager verifies every boot application binary against two lists: "db," which includes all signed and trusted programs, and "dbx," including all forbidden programs. But when a vulnerable binary is signed by Microsoft, the matter is moot.

Microsoft maintains a list of requirements for signing UEFI binaries, but the process is a bit obscure, Smolár says. "I don't know if it involves only running through this list of requirements, or if there are some other activities involved, like manual binary reviews where they look for not necessarily malicious, but insecure behavior," he says.

Microsoft has previously alluded to UEFI binaries being "approved through manual review." In response to an inquiry from Dark Reading, a Microsoft spokesperson provided the following clarification:

"Our process is designed to meet the evolving standard for analysis of third-party binary code, while ensuring effective scalability across our vast ecosystem. It begins with vetting and analysis with the partner, understanding the components they are requesting to be signed, and aspects of the design and functionality that may require additional security measures. Submissions are run through automated security analysis and are manually reviewed through a rigorous process. As a result, we may require additional security reviews, investigations with the submitting partner, or other mitigations. We remain committed to evolving our vetting process to better the security of our ecosystem as we operate within this ever-changing threat landscape."

ESET first discovered CVE-2024-7344 in July 2024. Since then, all vulnerable applications have been fixed, and Microsoft revoked the old, vulnerable binaries in its Jan. 14, 2025, Patch Tuesday update.

**About the Author**

Nate Nelson is a writer based in New York City. He formerly worked as a reporter at Threatpost, and wrote "Malicious Life," an award-winning Top 20 tech podcast on Apple and Spotify. Outside of Dark Reading, he also co-hosts "The Industrial Security Podcast."

Trusted Apps Sneak a Bug Into the UEFI Boot Process

Cybersecurity researchers have detailed an attack that involved a threat actor utilizing a Python-based backdoor to maintain persistent access to compromised endpoints and then leveraged this access to deploy the RansomHub ransomware throughout the target network.
According to GuidePoint Security, initial access is said to have been facilitated by means of a JavaScript malware downloaded named

Python-Based Malware Powers RansomHub Ransomware to Exploit Network Flaws

Ultimately, there is no replacement for an intuitive, security-focused developer working with the critical thinking required to drive down the risk of both AI and human error.

Matias Madou, Co-Founder & CTO, Secure Code Warrior

January 15, 2025

5 Min Read

Source: Nils Ackermann via Alamy Stock Vector

COMMENTARY

The advent of artificial intelligence (AI) coding tools undoubtedly signifies a new chapter in modern software development. With 63% of organizations currently piloting or deploying AI coding assistants into their development workflows, the genie is well and truly out of the bottle, and the industry must now make careful moves to integrate it as safely and efficiently as possible.

The OWASP Foundation has long been a champion of secure coding best practices, providing extensive coverage on how developers can best defend their codebases from exploitable vulnerabilities. Its recent update to the OWASP Top 10 for Large Language Model (LLM) Applications reveals the emerging and most potent threats perpetuated by AI-generated code and generative AI (GenAI) applications, and this is an essential starting point for understanding and mitigating the threats likely to rear their ugly heads. 

We must focus on integrating solid, foundational controls around developer risk management if we want to see more secure, higher quality software in the future, not to mention make a dent in the flurry of global guidelines that demand applications are released that are secure by design.

**The Perilous Crossover Between AI-Generated Code and Software Supply Chain Security**

Prompt Injection's ranking as the No. 1 entry on the latest OWASP Top 10 was unsurprising, given its function as a direct natural language command telling the software what to do (for better or worse). However, Supply Chain Vulnerabilities, which have a much more significant impact at the enterprise level, came in at No. 3. 

OWASP's advice mentions several attack vectors comprising this category of vulnerability, elements such as implementing pretrained models that are also precompromised with backdoors, malware and poisoned data, or vulnerable LoRA adapters that, ironically, are used to increase efficiency, but can, in turn, compromise the base LLM. These present potentially grave, widespread exploitable issues that can permeate the whole supply chain in which they are used.

Sadly, many developers are not skill- and process-enabled enough to navigate these problems safely, and this is even more apparent when assessing AI-generated code for business logic flaws. While not specifically listed as a category, as is apparent in OWASP’s Top 10 Web Application Security Risks, this is partly covered in No. 6, Excessive Agency. Often, a developer will vastly overprivilege the LLM for it to operate more seamlessly, especially in testing environments, or misinterpret how real users will interact with the software, leaving it vulnerable to exploitable logic bugs. These, too, affect supply chain applications and, overall, require a developer to apply critical thinking and threat modeling principles to overcome them. Unchecked AI tool use, or adding AI-powered layers to existing codebases, adds to the overall complexity and is a significant area of developer-driven risk.

**Data Exposure Is a Serious Concern Requiring Serious Awareness**

Sensitive Information Disclosure is second on the new list, but it should be a chief concern for enterprise security leaders and development managers. As OWASP points out, this vector can affect both the LLM itself and its application context, leading to personally identifiable information (PII) exposure, and disclosure of proprietary algorithms and business data.

The nature of how the technology operates can mean that exposing this data is as simple as using cunning prompts rather than actively "hacking" a code-level vulnerability, and "the grandma exploit" is a prime example of sensitive data being exposed due to lax security controls over executable prompts. Here, ChatGPT was duped into revealing the recipe for napalm when prompted to assume the role of a grandmother reading a bedtime story. A similar technique was also used to extract Windows 11 keys. 

Part of the reason this is made possible is through poorly configured model outputs that can expose proprietary training data, which can then be leveraged in inversion attacks to eventually circumvent the security controls. This is a high-risk area for those who are feeding training data into their own LLMs, and the use of the technology requires companywide, role-based security awareness upskilling. The developers building the platform must be well-versed in input validation and data sanitization (as in, these skills are verified and assessed before they can commit code), and every end user must be trained to avoid feeding sensitive data that can be spat out at a later date.

While this may seem trivial on a small scale, at the government or enterprise level, with the potential for tens of thousands of employees to inadvertently participate in exposing sensitive data, it's a significant expansion of an already unwieldy attack surface that must be addressed.

**Are You Paying Attention to Retrieval-Augmented Generation (RAG)?**

Perhaps the most notable new entry in the 2025 list is featured at No. 8, Vector and Embedding Weaknesses. With enterprise LLM applications often utilizing RAG technology as part of the software architecture, this is a vulnerability category to which the industry must pay close attention.

RAG is essential for model performance enhancement, often acting as the "glue" that provides contextual cues between pre-trained models and external knowledge sources. This is made possible by implementing vectors and embeddings, but if they are not implemented securely they can lead to disastrous data exposure, or pave the way for serious data poisoning and embedding inversion attacks. 

A comprehensive understanding of both core business logic and least-privilege access control should be considered a security skills baseline for developers working on internal models. However, realistically, the best-case scenario would involve utilizing the highest-performing, security-skilled developers and their AppSec counterparts to perform comprehensive threat modeling and ensure sufficient logging and monitoring.

As with all LLM technology, while this is a fascinating emerging space, it should be crafted and used with a high level of security knowledge and care. This list is a powerful, up-to-date foundation for the current threat landscape, but the environment will inevitably grow and change quickly. The way in which developers create applications is sure to be augmented in the next few years, but ultimately, there is no replacement for an intuitive, security-focused developer working with the critical thinking required to drive down the risk of both AI and human error.

**About the Author**

Co-Founder & CTO, Secure Code Warrior

Matias Madou is a researcher and developer with more than 15 years of hands-on software security experience. He has developed solutions for companies such as Fortify Software and his own company, Sensei Security. Over his career, Matias has led multiple application security research projects which have led to commercial products and boasts over 10 patents under his belt. When he is away from his desk, Matias has served as an instructor for advanced application security training courses and regularly speaks at global conferences, including RSA Conference, Black Hat, DEF CON, BSIMM, OWASP AppSec, and BruCon.

Matias holds a Ph.D. in Computer Engineering from Ghent University, where he studied application security through program obfuscation to hide the inner workings of an application.

OWASP's New LLM Top 10 Shows Emerging AI Threats

A critical vulnerability (CVE-2024-50603) in the Aviatrix Controller allows unauthenticated RCE. Active exploitation observed by Wiz Research in…

A critical vulnerability (CVE-2024-50603) in the Aviatrix Controller allows unauthenticated RCE. Active exploitation observed by Wiz Research in the wild for cryptojacking and backdoors. Learn about the risks and how to mitigate them.

Wiz Research, a prominent player in the cloud security space, has observed that a critical security flaw, CVE-2024-50603, impacting the Aviatrix Controller cloud networking platform, has been actively exploited in the wild by threat actors. With a CVSS score of 10.0, this critical vulnerability allows unauthenticated remote code execution (RCE) due to improper input sanitization in certain API endpoints.

CVE-2024-50603, is a critical flaw in Aviatrix Controller, a cloud networking platform, allowing unauthenticated remote code execution. A command injection vulnerability arises from improper input sanitization in certain API endpoints. That is, the Aviatrix Controller’s PHP API, which incorporates user-supplied parameters, is vulnerable to attacks due to improper handling, allowing malicious OS commands to be executed by unauthenticated users. 

What’s concerning is that attackers can exploit this flaw simply by crafting malicious commands. For example, instead of providing legitimate cloud-type values, an attacker could inject commands like: ; rm -rf / (to delete all files on the system), or, ;download\_malware.sh (to download and execute malicious software)

The issue arises from how these user-supplied parameters are integrated into the internal workings of the Aviatrix Controller. Instead of properly validating and sanitizing the input, these parameters are directly incorporated into system commands. Exploiting this flaw can grant attackers arbitrary command execution on the system, potentially leading to privilege escalation within the cloud environment.

Wiz Research has observed active exploitation of this vulnerability in the wild, with attackers deploying cryptocurrency miners and backdoors on compromised systems. This is concerning as the Aviatrix Controller often holds significant privileges within cloud environments, enabling attackers to move laterally and gain control of critical cloud resources. Given the widespread use of Aviatrix Controller and its potential for privilege escalation within cloud environments, this vulnerability can put the security of a wide range of organizations in danger.  

The vulnerability impacts Aviatrix Controller versions prior to 7.1.4191 and 7.2.4996. Security teams should immediately upgrade to the patched versions and implement network restrictions to minimize the attack surface, Wiz Research noted in the blog post shared with Hackread.com and published a proof-of-concept exploit, which can be accessed here. 

\[wp\_ad\_camp\_1

Proactive threat hunting is crucial, involving reviewing security logs for suspicious activity, searching for malware findings on compute resources hosting Aviatrix Controller, and analyzing cloud events for abnormal activity. By taking these steps, organizations can mitigate the risks associated with this critical vulnerability and prevent further exploitation.

Ray Kelly, a Fellow at Black Duck commented on the latest development emphasizing the importance of securing API endpoints. _“_The critical vulnerability (CVSS score: 10.0) in the Aviatrix Controller underscores the critical need to secure API endpoints. Developers often assume that APIs are hidden or immune to common web application attacks, but this example highlights how a server can be compromised through a simple web call._“_

_“_Thoroughly testing APIs is challenging due to their size, complexity, and the interdependence of chained calls. However, comprehensive security testing is essential, as neglecting it can lead to catastrophic consequences,_“_ Ray advised.

Hackers Use CVE-2024-50603 to Deploy Backdoor on Aviatrix Controllers

A critical security breach in the software supply chain has been detected. An attacker accessed Kong’s DockerHub account…

A critical security breach in the software supply chain has been detected. An attacker accessed Kong’s DockerHub account and replaced the legitimate Kong Ingress Controller v.3.4.0 image with a malicious version.

A critical security vulnerability was identified in Kong Ingress Controller version 3.4.0. This vulnerability stemmed from an unauthorized image uploaded to DockerHub on December 23rd, 2024. The affected image (hash: sha256:a00659df0771d076fc9d0baf1f2f45e81ec9f13179f499d4cd940f57afc75d43) contained malicious code that enabled cryptojacking. 

This malicious code directed the controller to pool.supportxmr.com, a crypto mining site, which means the image contained code that secretly mined cryptocurrency. Moreover, it would have been executed by any system running the compromised image, effectively turning those systems into unintended mining rigs.

The Kong team became aware of the issue on January 2nd, 2025 and took swift action. They removed version 3.4.0 and all associated tags from DockerHub. Additionally, they rotated all access keys used for DockerHub access. Finally, a patched version, 3.4.1, was released on January 2nd, 2025. This version removed the unauthorized cryptojacking code.

There is currently no evidence to suggest that any Kong Ingress Controller versions besides 3.4.0 (specifically the image hash mentioned above) were affected.

****What You Should Do:****

If you deployed Kong Ingress Controller version 3.4.0 between December 22nd, 2024 and January 3rd, 2025, immediate action is required. You should remove this image from all internal registries and clusters.

To fix the issue, remove the vulnerable image (sha256:a00659df0771d076fc9d0baf1f2f45e81ec9f13179f499d4cd940f57afc75d43) from internal registries and clusters, pull a remediated image from either the patched version 3.4.1 or a clean version 3.4.0.

The fixed image hashes for the clean, re-tagged version of 3.4.0 are:

AMD64: sha256:b358296fa6a1458c977c0513ff918e80b708fa9d7721f9d438f3dfce24f60f4f

ARM64: sha256:e0125aa85a4c9eef7822ba5234e90958c71e1d29474d6247adc3e7e21327e8ee

By taking these steps, you can ensure you are no longer running a vulnerable image and protect your systems from cryptojacking attempts.

Dan Lorenc, CEO and founder of software supply chain security platform Chainguard provided Hackread.com with a detailed comment addressing the core issues within this attack stating, “Supply-chain breaches happen, and it looks like the kong/kubernetes-ingress-controller images are the latest to fall victim. Here’s what we know so far:

*   A DockerHub PAT used to upload release images was compromised sometime before Dec. 23rd
*   The attacker used this PAT to upload a malicious version of the 3.4.0 release image directly to DockerHub
*   This image contained code to mine cryptocurrency and send results to a specific wallet
*   High CPU usage was reported by a user on December 29th, and the malicious images were taken down by January 2nd
*   New versions were uploaded, and the keys used to upload were revoked/rotated by the maintainers

“How should you protect against attacks like this? As a maintainer, any key you have is a key that you can leak,” said Dan. “Lock down and regularly audit all systems that have access to PATs like this, or choose systems that allow OIDC-based authentication to avoid this altogether. CI/CD pipelines are notoriously hard to configure securely; tools like Zizmor (lnkd.in/eGSGrMqm) help here. Signing artefacts can help even if you can’t use OIDC to publish or users pull from mirrors out of your control.”

“As an end-user, pin images you receive from third-parties by digest and test/malware scan them before upgrading. Check signatures if they exist,” Dan explained.

Nevertheless, a crypto mining attack on organizations could have drastic implications, including increased resource consumption, higher energy costs, and a multitude of security risks. The compromised image could have introduced vulnerabilities or backdoors, allowing attackers to gain further access, highlighting the importance of software supply chain security, particularly for critical components like container images. Organizations should employ image integrity verification mechanisms, and conduct regular security audits to identify and mitigate vulnerabilities.

1.  **OracleIV DDoS Botnet Hits Docker Engine API Instances**
2.  **Malware Hits 9Hits, Turns Docker Servers into Crypto Miners**
3.  **Hackers hijacking Bitbucket and Docker Hub for Monero mining**
4.  **TeamTNT Exploits 16M IPs in Malware Attack on Docker Clusters**
5.  **Linux Malware Alert: ‘Spinning YARN’ Hits Docker, Other Key Apps**

Tag

#backdoor