feat: replace server-side RAG with MCP retrieval primitives

- Remove Phase 4 RAG pipeline in favor of retrieval-only architecture
- Add FastMCP server exposing search, get_chunk, list_libraries tools
- Mount MCP endpoints (streamable HTTP + SSE) via Starlette in ASGI config
- Update README to clarify Mnemosyne is a retrieval engine, not RAG
- Let calling LLMs drive synthesis and iterative retrieval themselves

pyproject.toml

@@ -30,6 +30,9 @@ dependencies = [
     "tokenizers>=0.20,<1.0",
     "Pillow>=10.0,<12.0",
     "requests>=2.31,<3.0",
+    # Phase 5: MCP Server
+    "fastmcp>=2.0,<3.0",
+    "uvicorn[standard]>=0.30,<1.0",
 ]
 
 [project.optional-dependencies]