[Bug] Memory retrieval not applied to conversation context

Bug Description
Its just ignoring memories and not applying them to the current context - also doing a lot of obsessing over things that have been proven to be irrelevant.

Created a new feature /bug when I asked for ways to post a transcript to anthropic. Fessed up when challenged. I know a lot goes straight to the LLM but layering in for questions about that would be good

Also gets confused with the environment asks for feedback and processes the 1 2 3 4 feedback as if it were a request rather than feedback

Environment Info

• Platform: linux
• Terminal: pycharm
• Version: 2.1.126
• Feedback ID: 6bf2531c-a1a0-4443-8430-a1e5025bf9b0

Errors

[]