<p>This study explores secure communication methodologies that utilize multiple intelligent reflecting surfaces (IRSs) in millimeter-wave (mmWave) and terahertz (THz) communication frameworks. The investigation centers on a scenario in which a base station (BS) and a legitimate user (D) communicate exclusively via multiple IRSs, in the presence of an active eavesdropper (E). Notably, direct communication links between the BS and the user, as well as between the eavesdropper and the BS, are absent. Taking practical hardware implementation limitations into account, the research characterizes the conditions under which discrete phase shifts manifest in IRS reflections. Within this context, the principal aim is to maximize the secrecy rate of the mmWave/THz system. To tackle the ensuing non-convex optimization challenge, the problem is partitioned into two subproblems: first, the optimization of beamforming, and subsequently, the optimization of the reflection phase shift matrix. These subproblems are addressed in a joint optimization framework. A closed-form optimal solution is derived for the beamforming optimization, while for the reflection phase shift matrix optimization, a novel method integrating semidefinite relaxation (SDR) with element-wise block coordinate descent (E-BCD) is introduced. Simulation outcomes substantiate the efficacy of the proposed approaches.</p>