Warn , Limit or Prevent Potential LLM loops

[bashybaranaba]

LLMs enter a loop if you simply connect the resultant output of the world back to the LLM as an input.
We want to kinda allow users to be able to perform actions that call LLMs in a loop, but we need to have some sort of limits.

• First we want to warn world creators when their node structure is likely to result in a loop.
• If the user wants a loop, we need to have them specify the delay between the requests sent to the LLM bots.
• If a certain loop of events is too crazy just outright prevent it
• Also the token-based Onchain credit/token System  #148 system can be used to limit this behavior