feat(rag): add native Ollama RAG tool routing

[JerTheBear2018]

Summary

• Add native Ollama tool-calling support for RAG queries.
• Add manage_rag search support and Qwen tool-call parsing.
• Adapt NVIDIA GPU and native Ollama commands for the host setup.

Linked Issue

Fixes #5662

Type of Change

• Bug fix
• New feature
• Breaking change
• Documentation update

How to Test

1. Configure Odysseus to use native Ollama with qwen3:8b.
2. Ask: What do my notes say about the future NAS?
3. Confirm logs show tools_sent=7, a native call to manage_rag, and Tool executed: manage_rag: search.
4. Run:

python -m py_compile mcp_servers/rag_server.py src/ai_interaction.py src/tool_parsing.py
git diff --check

Checklist

• I searched existing issues and pull requests for duplicates.
• I tested these changes locally.
• The working tree is clean and the branch is rebased onto dev.

[JerTheBear2018]

All checks are passing. PR is ready for maintainer merge. Auto-merge is blocked because my account lacks EnablePullRequestAutoMerge permission.