Inside Ukraine’s Cyber Battle: How APT28 Breached Government Infrastructure with BEARDSHELL and SLIMAGENT

Listen to this Post

Featured Image
A Rising Threat: Ukraine Confronts a Sophisticated Cyber Offensive

In May 2025,

The attack combined social engineering, obfuscated delivery chains, registry hijacking, and abuse of trusted platforms like Signal, Icedrive, and Koofr. As cyber warfare escalates globally, this event serves as a critical case study in the importance of proactive threat detection, strong endpoint defenses, and the need for better controls over cloud service misuse.

Inside the Operation: APT28’s Infiltration of Ukrainian Government Systems

CERT-UA responded to an alarming breach in a central executive body’s communication infrastructure. Investigations revealed two potent malware tools—BEARDSHELL and SLIMAGENT—deployed on a compromised Windows server. BEARDSHELL is a C++ backdoor capable of decrypting and running PowerShell scripts while siphoning off data via Icedrive’s cloud API. It uniquely assigns a hashed directory to each infected host, improving stealth and traceability. Meanwhile, SLIMAGENT, also written in C++, performed covert reconnaissance, capturing screenshots using Windows GDI APIs and encrypting them with AES and RSA before local storage.

The attackers’ persistence was strategic. Although the initial entry point was unclear, a follow-up campaign in May shed light on the method: a booby-trapped document named “Act.doc” sent through the Signal messaging app. Macros embedded within the document, once enabled by unsuspecting users, executed shellcode that pulled in malware through COM-hijacking and loaded components of the COVENANT framework, an open-source remote control tool using the Koofr platform for C2 operations.

The infection chain included multiple payloads like “PlaySndSrv.dll” and “sample-03.wav”, the latter containing encrypted code to reactivate BEARDSHELL. The attack ensured resilience through Windows Registry tweaks and scheduled tasks tied to system sound services, enabling malware persistence even after reboots.

CERT-UA, supported by military cyber unit A0334 and ESET researchers, discovered that both BEARDSHELL and SLIMAGENT used legitimate cloud platforms for control communications, bypassing conventional network monitoring. APT28’s playbook now includes manipulating multimedia features, registry COM classes, and trusted storage domains for deep intrusion and covert surveillance.

As part of its response, CERT-UA advised organizations to monitor traffic directed to app.koofr.net and api.icedrive.net. Their analysis also emphasized tightening macro execution policies, enhancing endpoint detection, and scrutinizing Signal-based communications.

The entire operation reflects a significant evolution in cyber espionage—merging technical depth with psychological tactics and exploiting common weaknesses in endpoint and user behavior. It’s a stark reminder that even state-level digital infrastructures are vulnerable without adaptive, threat-informed defense strategies.

What Undercode Say: Dissecting the Strategic Depth of APT28’s Attack

APT28’s evolving methods mirror modern hybrid warfare. The

The dual use of BEARDSHELL and SLIMAGENT reveals tactical layering. While BEARDSHELL handled deeper system control through PowerShell, SLIMAGENT maintained surveillance—two distinct roles for maximum intelligence gathering and persistence. The use of ChaCha20-Poly1305, AES, and RSA encryption suggests high operational security. These were not smash-and-grab tools; they were designed for stealth, endurance, and data theft.

Signal as a delivery vector adds a novel twist. Messaging apps like Signal are typically trusted environments, but that trust can be weaponized. The use of Signal to distribute “Act.doc” points to high adversary awareness of operational security procedures and a deep understanding of their targets’ internal workflows. This human-centric angle complements their technical sophistication.

Macro-based payloads remain an evergreen threat. Despite years of awareness campaigns, macros are still an effective entry point. APT28 leveraged macro-enabled documents to stage shellcode loaders and pivot to further downloads. Organizations still struggle to control macro execution, especially when users operate with elevated privileges or lax controls.

COVENANT’s use as a modular framework shows open-source risks. The inclusion of the COVENANT framework, freely available online, shows how attackers blend community tools with custom malware. This modularity helps them evade attribution and iterate quickly across targets.

Registry COM hijacking and sound service abuse mark creative persistence techniques. Attackers exploited CLSID registry paths and tied scheduled tasks to Windows’ multimedia features. This shows an in-depth understanding of Windows internals and a desire to make detection difficult even for well-prepared blue teams.

CERT-UA’s collaboration with cyber-military units shows a strong defensive model. The joint effort between CERT-UA, military units, and private sector vendors like ESET exemplifies a mature incident response pipeline. Rapid sharing of IOCs and behavioral patterns can shorten the attacker’s dwell time and improve defenses across other agencies.

The attack’s broader implications point to escalating cyber hostilities. APT28’s use of espionage-grade tactics in peacetime indicates that digital infrastructure is now a permanent theater for geopolitical conflict. Governments worldwide must reassess their defenses not just against generic malware, but against tailored, persistent, and cloud-leveraging intrusions.

For defenders, the lessons are sobering. Macros must be disabled by default. Messaging apps should be sandboxed or monitored for unusual file transfers. Legitimate domains like Icedrive and Koofr need behavioral analytics layered on top of simple allowlists. Organizations should deploy threat hunting teams with specific focuses on registry changes, screenshot-capturing APIs, and scheduled task abuse.

APT28’s innovation won’t stop here. Future campaigns may include LLM-powered phishing, AI-based evasion, and expanded use of encrypted DNS. Defenders must adapt, leveraging automation, AI detection, and threat intelligence to stay ahead.

🔍 Fact Checker Results:

✅ BEARDSHELL and SLIMAGENT were confirmed by CERT-UA and ESET as part of the May 2025 campaign
✅ APT28’s involvement is corroborated through behavioral and technical indicators
✅ Abuse of Icedrive, Koofr, and Signal platforms has been publicly validated through threat reports

📊 Prediction:

🔮 APT28 and similar state-sponsored groups will likely increase their reliance on cloud services and messaging apps to conduct stealth operations.
🔮 Expect deeper malware integration into legitimate software components, especially Windows services tied to media or productivity.
🔮 Future campaigns may blur the lines between cybercrime and cyberwarfare, targeting both civilian infrastructure and military-adjacent systems.

References:

Reported By: cyberpress.org
Extra Source Hub:
https://stackoverflow.com
Wikipedia
OpenAi & Undercode AI

Image Source:

Unsplash
Undercode AI DI v2

Join Our Cyber World:

💬 Whatsapp | 💬 Telegram