Governmental and military institutions harbor critical infrastructure and highly confidential information. Although institutions are investing a lot for protecting their data and assets from possible outsider attacks, insiders are still a distrustful source of information leakage. As malicious software injection is one among many attacks, turning innocent employees into malicious attackers through social attacks is the most impactful one. Malicious insiders or uneducated employees are dangerous for organizations that they are already behind the perimeter protections that guard the digital assets; actually, they are trojans on their own. For an insider, the easiest possible way for creating a hole in security is using the popular and ubiquitous Universal Serial Bus (USB) devices due to its versatile and easy to use plug-and-play nature. USB type storage devices are the biggest threats for contaminating mission critical infrastructure with viruses, malware, and trojans. USB human interface devices are also dangerous as they may connect to a host with destructive hidden functionalities. In this paper, we propose a novel hardware-assisted insider threat detection and prevention framework for the USB case. Our novel framework is also OS independent. We implemented a proof-of-concept design on an FPGA board which is widely used in military settings supporting critical missions, and demonstrated the results considering different experiments. Based on the results of these experiments, we show that our framework can identify rapid-keyboard key-stroke attacks and can easily detect the functionality of the USB device plugged in. We present the resource consumption of our framework on the FPGA for its utilization on a host controller device. We show that our hard-to-tamper framework introduces no overhead in USB communication in terms of user experience.