The Val d'Agri region (Southern Italy) hosts high-rate, diffuse seismicity, that has been attributed to the complex interplay of tectonic and anthropogenic processes. In this work we apply waveform similarity-based clustering, earthquake relocation, and moment tensor inversion to a newly developed high-resolution microseismic catalog. This combined approach allows us to image the geometry of active seismogenic structures and characterize their faulting style. Waveform similarity-based clustering identifies groups of earthquakes having highly similar waveforms, interpreted as likely associated with the same seismogenic structures. High-quality relocation of these events allows the delineation of planar features, suggestive of fault segments.The analysis of the largest events (M > 2) by moment tensor inversion provides constraints on the prevailing faulting style. Our results indicate that the seismicity in the analyzed time period predominantly occurs along steeply southwest-dipping seismogenic structures, with moment tensor solutions consistent with the regional extensional stress regime. We further analyze a seismic sequence occurred during the study timespan to investigate its spatio-temporal evolution and possible controlling processes. We explore scenario-based rupture models for the largest imaged structure using finite-fault ground-motion simulations, providing first-order physics-based constraints on potential ground-motion variability.