Sorry, the LLM returned a bad or malformed response. If this persists,
try clearing Cliff's memory with "cliff --memory clear", and if that
still doesn't work, try switching to a different model.
