This is pretty neat. We appear to be working on projects in parallel. It seems like Maelstrom focuses on the decision-making under stress conceptually, while my persistent agent framework works on learning and correcting in real time, pretty neat to see the adjacent problem being worked on.
