Mitigating Bias in Conversations: A Hate Speech Classifier and Debiaser with Prompts