Attack graphs are often utilized by researchers for analysis of a network security posture. However, available implementations of attack graph generation tools are either terribly structured and chaotic or the source code is not available. Therefore custom modification for research purposes is not possible. The goal of the thesis is to implement an algorithm for attack graph generation as described in enclosed literature and extend the approach by considering more granular vulnerability specification. The work will consist of following tasks:

• implementation of preprocessor, which will convert the JSON configuration input to rules, including network topology, vulnerability descriptions, and filtering information,
• implementation of an algorithm for attack graph generation based on output from previous task,
• representation of the algorithm output in JSON format,
• analysis of algorithm's time complexity.

The results will be utilized in project CRUSOE - Research of Tools for Cyber Situational Awareness and Decision Support of CSIRT Teams in Protection of Critical Infrastructure.