With the introduction of SysML v2—featuring a new metamodel and standardized APIs—new opportunities arise for automating MBSE tasks through generative AI. However, existing LLM approaches lack interoperability due to limited support for complex SysML v2 API operations. This paper presents a framework for implementing Model Context Protocol (MCP) servers for SysML v2, enabling LLMs to interact with system models via the API. This framework provides a foundation for agentic MBSE by allowing automation of modeling tasks while maintaining interoperability across tools and generative AI systems.
