Modeling Penetration Testing with Reinforcement Learning Using Capture-the-Flag Challenges and Tabular Q-Learning