Domain-Grounded Tool Orchestration for LLM-Guided Scientific Analysis
This work presents an architecture that separates intent interpretation (LLM) from execution (deterministic domain tools) from explanation (LLM), connected by the Model Context Protocol (MCP) and grounded by domain ontologies that constrain planning to valid analysis chains.
Jeff Lee, S. Jourdain, Cory Quammen et al.
· 0 citations