Integrate RAG for both background as well as for knowledge-database #4
Description
After I have stabilized the state-data flows, I want to investigate usage of RAG within my prototype
- for storing background knowledge (e.g., hacking manuals)
- for dynamically storing and accessing gathered knowledge about the observed system
If I integrate RAG within the high-level planner as tooling, it might work.. as long as the LLM/planner knows that it should query RAG for background knowledge. Still need to think how to best to integrate this (RAG through context, or calling RAG operations as tooling).