Conversational recommender systems (CRS) have advanced with large language models, showing strong results in domains like movies. These domains typically involve fixed content and passive consumption, where user preferences can be matched by genre or theme. In contrast, games present distinct challenges: fast-evolving catalogs, interaction-driven preferences (e.g., skill level, mechanics, hardware), and increased risk of unsafe responses in open-ended conversation. We propose MATCHA, a multi-agent framework for CRS that assigns specialized agents for intent parsing, tool-augmented retrieval, multi-LLM ranking with reflection, explanation, and risk control which enabling finer personalization, long-tail coverage, and stronger safety. Evaluated on real user request dataset, MATCHA outperforms six baselines across eight metrics, improving Hit@5 by 20%, reducing popularity bias by 24%, and achieving 97.9% adversarial defense. Human and virtual-judge evaluations confirm improved explanation quality and user alignment.

翻译：对话式推荐系统（CRS）借助大语言模型取得了进展，在电影等领域展现出强劲的性能。这些领域通常涉及固定内容和被动消费，用户偏好可以通过类型或主题进行匹配。相比之下，游戏领域呈现出独特的挑战：快速更新的产品目录、由交互驱动的偏好（例如技能水平、操作机制、硬件配置），以及在开放式对话中更高的不安全回复风险。我们提出了MATCHA，一个用于CRS的多智能体框架，它分配了专门的智能体来处理意图解析、工具增强检索、带反思的多LLM排序、解释生成和风险控制，从而实现更精细的个性化、更好的长尾覆盖和更强的安全性。在真实用户请求数据集上的评估表明，MATCHA在八项指标上均优于六个基线模型，将Hit@5提升了20%，将流行度偏差降低了24%，并实现了97.9%的对抗性防御能力。人工和虚拟评估均证实了其解释质量的提升和与用户期望的更好对齐。