Hypothesis storage #4

Void's hypothesis mechanism currently works as follows:

1. Track working hypotheses in a memory block
2. Request or find information related to those hypotheses until they can be closed
3. Migrate finished hypotheses to archival memory

Arguably, hypotheses should be treated separately. My proposal would be to create a specific Letta memory block with a full report, and then simply detach the block. This would allow recall of a specific hypothesis report rather than relying on archival search methods.