Anomaly detection systems must be highly resilient to safeguard computer networks against the growing threat of network attacks. Security vulnerabilities are inevitable when conventional approaches fail to adapt to the constantly evolving realm of cyber threats. This study presents VigilNet, an advanced anomaly detection framework that combines supervised learning techniques. VigilNet improves its capability to detect and withstand various network intrusions by utilizing a Voting Classifier ensemble that combines the Random Forest and K-Nearest Neighbors (KNN) classifiers. VigilNet addresses the issue of limited labeled training data for network security applications by incorporating both unlabeled and labeled data, effectively overcoming this challenge. Research findings suggest that VigilNet is a highly effective tool for achieving precise and robust performance. Due to its capabilities, it has the potential to address the ever-expanding threat landscape and bring about a significant transformation in how network security detects anomalies.