Semantic / Hybrid Search using MCP server

Be able to index vectors using an embedding model in Solr and perform similarity search as well as hybrid (lexical + semantic) search.

Start by reviewing the Solr MCP server's existing search entry points and how it communicates with Solr. Determine how embedding-based similarity search and combined lexical-plus-semantic search should be exposed, then define completion criteria for indexing vectors and returning both result types.

Written by the indexing model from the issue text.