New experimental prompt strategy to only show valid and invalid combinations of behavior emojis, instead of example phrases. Works well with gpt-4o-mini and llama3-70b. Does not work great with gpt-3.5-turbo. Also added negative LLM unit tests checks for behaviors (FOLLOW and not ATTACK).

- new junit test module
- 4 initial tests for attack, follow, and flee
- improved regex to support <behavior> or *behavior*
- improved message cleaning
- refactor of HTTP requests to remove some Minecraft and Fabric specific imports