LLM-based agents have emerged as innovative tools capable of performing complex tasks through iterative planning and action, making significant advances in understanding and addressing user needs. However, their effectiveness is limited in specialized areas such as mental health diagnosis, as they underperform compared to general applications. Current approaches to integrating diagnostic capabilities into LLM rely on rare and highly sensitive mental health datasets, which are difficult to obtain. Furthermore, these methods fail to mimic clinicians' pre-questioning skills, lack multi-session conversational comprehension, and struggle to align their results with expert clinical reasoning. To address these challenges, this study proposes DSM5AgentFlow, the first LLM-based agent workflow designed to autonomously generate DSM-5 Level-1 diagnostic questionnaires. By simulating therapist-client conversations with specific client profiles, this framework provides transparent, step-by-step disorder predictions, generating explainable and reliable results. This workflow serves as a complementary tool for mental health diagnosis that adheres to ethical and legal standards. Through comprehensive experiments, we evaluate key LLMs across three key dimensions: conversational realism, diagnostic accuracy, and explainability. The dataset and implementation are fully open source.
