This paper received the Best Graduate Forum Paper (Runner-Up) award at IEEE COMSNETS 2026.
The integration of Large Language Models (LLMs) into network operations (AIOps) is hindered by two fundamental challenges: the stochastic grounding problem, where LLMs struggle to reliably parse unstructured, vendor-specific CLI output, and the security gap of granting autonomous agents shell access. This paper introduces MCP-Diag, a hybrid neuro-symbolic architecture built upon the Model Context Protocol (MCP). We propose a deterministic translation layer that converts raw stdout from canonical utilities (dig, ping, traceroute) into rigorous JSON schemas before AI ingestion. We further introduce a mandatory "Elicitation Loop" that enforces Human-in-the-Loop (HITL) authorization at the protocol level. Our preliminary evaluation demonstrates that MCP-Diag achieving 100% entity extraction accuracy with less than 0.9% execution latency overhead and 3.7x increase in context token usage.
@misc{lodha2026mcpdiag,title={MCP-Diag: A Deterministic, Protocol-Driven Architecture for AI-Native Network Diagnostics},author={Lodha, Devansh and Panchal, Mohit and Kulkarni, Sameer G.},year={2026},archiveprefix={arXiv},primaryclass={cs.NI},url={https://arxiv.org/abs/2601.22633},}
MCP-Diag: A Deterministic, Protocol-Driven Architecture for AI-Native Network Diagnostics2026