The Pakistan-aligned threat group Transparent Tribe (also known as APT36) has begun leveraging artificial intelligence-assisted coding tools to mass-produce malware implants in a large-scale cyber-espionage campaign primarily targeting India. According to research from Bitdefender, the attackers are not focusing on highly sophisticated malware but instead generating a large volume of relatively simple, disposable binaries using less common programming languages such as Nim, Zig, and Crystal. This strategy aims to overwhelm detection systems by creating diverse malware variants that communicate using different protocols. Researchers describe this tactic as a Distributed Denial of Detection (DDoD) approach, where threat actors flood environments with numerous malware samples to complicate security monitoring and analysis. The rise of large language models has significantly reduced the technical barrier for attackers, enabling them to rapidly generate functional malicious code in unfamiliar languages. The campaign specifically targets Indian government entities, diplomatic missions abroad, and to a lesser extent organizations linked to the Afghan government and private sector businesses. The attackers reportedly identify high-value targets through platforms such as LinkedIn. Infection chains typically begin with phishing emails containing ZIP or ISO archives that include malicious Windows shortcut (LNK) files. In other cases, victims receive PDF documents containing a “Download Document” lure that redirects them to a malicious website hosting the same archives. Once executed, the LNK files launch PowerShell scripts that run in memory to download the main backdoor and additional payloads. The attackers also deploy well-known post-exploitation frameworks such as Cobalt Strike and Havoc C2 Framework to maintain persistence and enable further compromise. The operation includes numerous custom malware tools such as Warcode, NimShellcodeLoader, SupaServ, LuminousStealer, CrystalShell, ZigShell, and BackupSpy. Many of these tools exploit legitimate cloud services like Slack, Discord, Google services (including Google Sheets and Google Drive), and Supabase for command-and-control communication and data exfiltration. Although AI-assisted malware development allows attackers to rapidly scale their campaigns, researchers note that many of the generated tools are unstable and contain coding errors. Nevertheless, the massive volume of implants combined with trusted infrastructure and varied programming languages makes detection more challenging for organizations and increases the overall operational impact of the campaign.
Linux rootkits have become a serious threat to modern infrastructure as Linux adoption grows across cloud platforms, container environments, IoT devices, and high-performance serve...
Security researchers have uncovered a sophisticated malware campaign named VOID#GEIST, which uses a multi-stage delivery framework to deploy several remote access trojans (RATs). T...
Cisco has confirmed that two vulnerabilities affecting Catalyst SD-WAN Manager are being actively exploited in real-world attacks. The flaws, tracked as CVE-2026-20128 and CVE-2026...