The phage lysin field has done nothing but grow in the last decades. As a result, many different research groups around the world are contributing to the field, often with certain methodological differences that pose a challenge to the interpretation and comparison of results. In this work, we present the case study of three Acinetobacter baumannii-targeting phage lysins (wild-type endolysin LysMK34 plus engineered lysins eLysMK34 and 1D10) plus one lysin with broad activity against Gram-positive bacteria (PlySs2) to provide exemplary evidence on the risks of generalization when using one of the most common lysin evaluation assays: the killing assay with resting cells. To that end, we performed killing assays with the aforementioned lysins using hypo-, iso- and hypertonic buffers plus human serum either as the reaction or the dilution medium in a systematic manner. Our findings stress the perils of creating hypotonic conditions or a hypotonic shock during a killing assay, suggesting that hypotonic buffers should be avoided as a test environment or as diluents before plating to avoid overestimation of the killing effect in the assayed conditions. As a conclusion, we suggest that the nature of both the incubation and the dilution buffers should be always clearly identified when reporting killing activity data, and that for experimental consistency the same incubation buffer should be used as a diluent for posterior serial dilution and plating unless explicitly required by the experimental design. In addition, the most appropriate buffer mimicking the final application must be chosen to obtain relevant results.