TrustZone-based Trusted Execution Environments (TEEs) have been utilized extensively for the implementation of security-oriented solutions for several smart intra and inter-connected devices. Although TEEs have been promoted as the starting point for establishing a device root of trust, a number of published attacks against the most broadly utilized TEE implementations request a second view on their security. The aim of this research is to provide an analytical and educational exploration of TrustZone-based TEE vulnerabilities with the goal of pinpointing design and implementation flaws. To this end, we provide a taxonomy of TrustZone attacks, analyze them, and more importantly derive a set of critical observations regarding their nature. We perform a critical appraisal of the vulnerabilities to shed light on their underlying causes and we deduce that their manifestation is the joint effect of several parameters that lead to this situation. The most important ones are the closed implementations, the lack of security mechanisms, the shared resource architecture, and the absence of tools to audit trusted applications. Finally, given the severity of the identified issues, we propose possible improvements that could be adopted by TEE implementers to remedy and improve the security posture of TrustZone and future research directions.