Featured · ML-driven IDS
NetInfect
Machine-learning intrusion detection with real-time active response, running on an emulated three-zone enterprise network.
- Kathará + Docker lab with WAN, DMZ and LAN zones, and Zeek on the gateway
- Hybrid model: an Isolation Forest anomaly score feeds a Random Forest classifier
- Scripted attacks for labelled data: port scan, SSH brute force, SQLi/XSS, HTTP flood, Slowloris, exfiltration
- Active response: iptables rules injected live, regex-based L7 WAF, Streamlit dashboard
- Python
- scikit-learn
- Zeek
- Kathará
- iptables
- network trafficWAN → gateway
- Zeekconn.log · http.log
- Isolation Forestanomaly score
- Random Forestattack / benign
- iptables DROPattacker isolated