This document presents a model for detecting the agent responsible for data leakage. It discusses adding fake objects to distributed data in order to identify the source if a leakage occurs. The model is implemented using C# and SQL Server. When an agent requests data, the distributor sends the original data along with randomly allocated fake objects. If the data is leaked, the distributor can analyze the fake objects to determine the guilty agent. An algorithm is provided and screenshots show modules for login, data sharing, and detecting the guilty agent using a probability calculation. The model aims to overcome limitations of existing watermarking techniques for data leakage detection.