The effort to model and understand knowledge and data has given rise to a large variety of implementations of knowledge level modeling. 
Aimed at achieving various operations such as anomaly detection, classification and planning, among others, knowledge level modeling allows us to derive generalizations concerning data. 
In this paper, we present a toolkit aimed conducting KL modeling. 
This toolkit will be presented at this stage of its implementation to allow for anomaly detection in classification data.
By using this toolkit, we implemented a two-step, likelihood-based anomaly detector and tested it on simulated classification datasets for various different scenarios.
Our model has achieved to perfectly identify the normal and abnormal test instances from the simulated datasets.