Cybersecurity researchers have reported that the North Korean-linked hacking group known as Konni, also tracked as Opal Sleet or TA406, has begun deploying malware that shows signs of being developed with the assistance of artificial intelligence. The campaign has been observed targeting developers and engineers working in the blockchain and cryptocurrency sector, signaling a continued focus on high-value digital assets and infrastructure.
Konni is widely believed to be connected to other North Korean activity clusters such as APT37 and Kimsuky. The group has been active for more than a decade and has previously focused its operations on organizations across South Korea, Russia, Ukraine, and multiple European countries. Recent findings indicate that its operational scope now places particular emphasis on the Asia-Pacific region.
Asia-Pacific Focus and Initial Infection Vector
Analysis conducted by Check Point researchers revealed that samples linked to the latest campaign were submitted from countries including Japan, Australia, and India. This geographic spread suggests a deliberate effort to compromise targets within the broader Asia-Pacific technology ecosystem.
The infection chain begins with a social engineering approach. Victims receive a link hosted on Discord that leads to the download of a ZIP archive. This archive contains a decoy PDF file alongside a malicious shortcut file in LNK format. When the shortcut is opened, it triggers an embedded PowerShell loader that extracts additional components onto the system.
These components include a DOCX document used as a lure, a cabinet archive, and multiple malicious elements such as a PowerShell backdoor, two batch scripts, and an executable designed to bypass User Account Control protections.
Compromising Development Environments
Opening the shortcut file results in the DOCX document being displayed to the user while, in the background, one of the batch files is executed. According to researchers, the content of the lure document points to a strategic goal of infiltrating development environments. Gaining access at this level could allow attackers to reach critical assets such as infrastructure configurations, application programming interface credentials, digital wallets, and ultimately cryptocurrency funds.
One of the batch scripts creates a staging directory for the backdoor and associated files, while the second establishes a scheduled task that runs hourly. This task is disguised as a legitimate OneDrive startup process, helping it blend into normal system activity.
Obfuscation and AI-Assisted Code Indicators
The scheduled task reads an encrypted PowerShell script from disk, decrypts it using an XOR routine, and executes it directly in memory. After execution, the task removes itself in an effort to erase evidence of compromise. The backdoor code itself is heavily obfuscated through arithmetic-based string encoding, dynamic string reconstruction at runtime, and final execution using PowerShell expression invocation.
Check Point researchers assessed that several characteristics of the script strongly suggested AI-assisted development rather than malware written entirely by a human operator. Indicators included unusually clear and structured documentation within the script, a clean and modular layout, and comments that resembled instructional placeholders commonly seen in large language model outputs and coding tutorials.
Command-and-Control and Attribution
Before fully activating, the malware conducts checks of hardware, software, and user behavior to confirm it is not operating in a sandbox or analysis environment. It then generates a unique identifier for the infected host. Depending on the level of privileges available, the backdoor follows different execution paths.
Once active, the malware periodically communicates with its command-and-control server, transmitting basic system metadata and polling for further instructions at randomized intervals. If the server responds with additional PowerShell code, the backdoor executes it asynchronously using background jobs.
Researchers attributed the campaign to Konni based on similarities with previous attacks, including overlapping launcher formats, lure file naming patterns, and consistent execution chain structures. To assist defenders, Check Point has released indicators of compromise associated with the campaign to help organizations detect and mitigate potential infections.