Abstract: A method and device are disclosed for automatically evaluating a software source code quality, wherein evaluation rules and/or metrics for evaluating the software source code are pre-set. The source code is inspected with the aid of a set of evaluation rules and/or metrics, wherein the set contains at least one part of pre-set evaluation rules and/or metrics. In addition, in at lest one embodiment, the set of evaluation rules and/or metrics used for inspecting the source code is adapted according to the evaluation of a performed inspection of the source code with respect to at lest one predefined criterion for forming an adapted set of evaluation rules and/or metrics different from the first set. In addition, the source code is inspected by way of the adapted set of evaluation rules and/or metrics.