[Question]: Unable to refer to retrieval result

[mighty-laolin]

Describe your problem

Here is how the canvas look like:

Basically I would like the agent to first do a retrieval, and then allow the user to make conversations based on those retrieved documents.

However, the retrieved documents are no longer visible to the model in the 2nd round conversation onward. Not sure if this is by design or a bug, but it would definitely be helpful to access whatever has been retrieved when having multi-round conversation.

[KevinHuSh]

[mighty-laolin]

Tried this already. This way it does a retrieval every time and the conversation context is lost somehow.

[mighty-laolin]