Анализ структуры MPS-языка и его концепций

★ 7.9 · vibe-coding

mps-language-analysis is a Claude Code skill that inspects an MPS language by its fully qualified name, discovering concepts, properties, references, children, aspects (editor, constraints, behavior), and metadata. The workflow chains several MCP calls: `mps_mcp_get_project_structure` to confirm the language and retrieve its UUID, `mps_mcp_get_concept_details` to list concepts with their FQNs and structural fields, `mps_mcp_query_nodes` with `FIND_INSTANCES` to locate sample nodes, and `mps_mcp_query_structure` with `LIST_CONCEPT_ASPECTS` to surface associated editor and behavior definitions. Concept metadata — `isRootable`, `isAbstract`, and `conceptReference` IDs — feeds directly into JSON blueprints used when creating or modifying nodes. The skill is aimed at developers exploring unfamiliar MPS languages or reverse-engineering concept hierarchies before authoring new language extensions. Related skills `mps-language-inheritance` and `mps-language-aspects-overview` extend the analysis into superconcept trees and aspect ownership.