Security researchers have identified a new Android malware framework dubbed Bazuka, which leverages sophisticated screen overlay techniques to facilitate credential theft and unauthorized device control. Discovered by mobile security firm ThreatFabric during an investigation into NFC-relay banking Trojans, the malware is notable for its integration of large language model (LLM) capabilities to automate and scale the management of compromised devices.
The primary mechanism employed by Bazuka involves the deployment of deceptive screen overlays that mimic legitimate banking or service applications. By drawing these overlays over genuine interfaces, the malware can intercept user inputs, including login credentials and two-factor authentication codes, without the user realizing they are interacting with a malicious layer. This technique effectively bypasses standard visual verification by the victim, allowing the attackers to harvest sensitive data in real time.
Beyond simple data exfiltration, the framework is designed to grant attackers persistent control over the infected device. By abusing Android’s accessibility services, the malware can simulate user interactions, navigate menus, and execute commands as if the device owner were operating the phone. The addition of LLM integration suggests a shift toward more automated, intelligent victim management, potentially allowing operators to handle a larger volume of compromised devices with reduced manual effort.
Products in this category commonly target the Android operating system by requesting excessive permissions during installation, particularly those related to accessibility and screen drawing. When these permissions are granted, they provide the malware with the necessary privileges to monitor activity, capture keystrokes, and manipulate the user interface. Mitigation for this class of threat typically involves restricting the installation of applications to official, verified sources and strictly auditing the permission requests of any app that requires accessibility service access.
The emergence of Bazuka highlights a broader trend in mobile threat development where modularity and automation are becoming standard features of malware frameworks. As attackers incorporate advanced technologies like LLMs into their toolkits, the barrier to entry for managing large-scale botnets is lowered, and the efficacy of social engineering and credential harvesting campaigns is significantly enhanced. This development underscores the ongoing challenge of securing mobile environments against increasingly sophisticated and automated exploitation techniques.





