AgenticNet: Utilizing AI Coding Agents To Create Hybrid Network Experiments

Collection of recent network and AI-for-networking papers with mixed security implications. Key risks include automated policy generation via LLMs (Intent-Based Networking) and AI coding agents that can introduce misconfiguration, conflict, or unsafe network changes; E2E orchestration and network-slicing platforms that enlarge attack surface and risk cross-slice isolation failures; and observable GenAI traffic fingerprints (SNI/QUIC/TLS patterns) enabling traffic classification, privacy leakage, or exfiltration. Positive security advances include integration of QKD with IPsec for quantum-safe,