🧠 Issue 12: Add RAG-based Generative AI interface to Sztab

Description: Prototype an interface allowing a non-technical user to interact with Sztab using natural language.

• Use a lightweight LLM (e.g., Claude, LLaMA) connected via RAG
• REST API responses from Sztab serve as the context source

Tasks:

•  Define a set of representative queries (e.g. overdue issues, project timelines)
•  Set up a small vector DB or context indexer for Sztab data
•  Build RAG prototype:
  • LLM frontend
  • Retriever backend (REST calls to Sztab)
•  Integrate responses via chat or CLI (not necessarily ChatGPT UI)
•  Evaluate Anthropic/Claude or LLaMA for local development

Estimated Effort: 16–20 hours Milestone: Milestone 2 – Post-MVP