Reconfigurable intelligent surface (RIS) has been envisioned as a key technology in future wireless communication networks to enable smart radio environment. To further enhance the passive beamforming capability of RIS, beyond diagonal (BD)-RIS has been proposed considering interconnections among different RIS elements. BD-RIS has a unique feature that cannot be enabled by conventional diagonal RIS; it can be realized by non-reciprocal circuits and thus has asymmetric scattering matrix. This feature provides probability to break the wireless channel reciprocity, and thus has potential to benefit the full-duplex (FD) system. In this paper, we model the BD RIS-assisted FD systems, where the impact of BD-RIS non-reciprocity and that of structural scattering, which refers to the virtual direct channel constructed by RIS when the RIS is turned OFF, are explicitly captured. To visualize the analysis, we propose to design the scattering matrix, precoder and combiner to maximize the DL and UL sum-rates in the FD system. To tackle this optimization problem, we propose an iterative algorithm based on block coordination descent (BCD) and penalty dual decomposition (PDD). Numerical results demonstrate surprising benefits of non-reciprocal BD-RIS that it can achieve higher DL and UL sum-rates in the FD scenario than reciprocal BD-RIS and conventional diagonal RIS.