Story

As a developer of RHDH, I want to make sure that the deep context retrieval is something we can achieve technically so that the implementation phase of this feature can begin.

Background

This investigation is about the technical aspect of the deep context awareness feature. Our main focus will be the backend implementation.

That said we're interesting in investigating if a solution around the Deep Context Retrieval is feasible.

Here is important to highlight all the limitations, as well as, highlight any extra capabilities. Something we could explore for example as part of this investigation is also:

Could we combine the deep context awareness with MCP actions. E.g when the LLM agent understands that it’s in a position to perform an action (i.e. create a catalog entity) should it perform it? And if yes how?

Dependencies and Blockers

QE impacted work

Documentation impacted work

Acceptance Criteria

• A report has been created for the backend approach of the deep context retrieval feature.

upstream documentation updates (design docs, release notes etc)

Technical enablement / Demo