Manual composition of tasks and exams is a challenging and time-consuming task. Especially when exams are taken remotely without the personal monitoring by examiners, most exams can easily lose their integrity with the use of previously done exercises or student communication. This research introduces an approach that incorporates the principles of the generative software development and aspects of the feature-oriented product line engineering process into the field of question creation and generation. The resulting generator can be used to generate single-choice-question-families by means of written question templates. The generated questions within a question-family differ based on the set features and parameters and can be imported into the target learning management system ILIAS. Without much effort, examiners and educators can use the generator to create variants of their questions and deliver them to their students.