AI 中文总结
研究针对代码理解中LLM助手的不足，提出TARS智能体，基于心理理论范式，剖析开发者特点并调整解释，通过检索增强生成联系项目文档。实验表明它能提升效率、降低认知负荷，解释更适配开发者。
AI 中文摘要
代码理解是软件工程中最耗时的任务之一，然而大多数基于大语言模型（LLM）的助手给出的解释忽略了提问者，还迫使开发者采用复制粘贴的繁琐工作流程。我们展示了TARS，这是一个集成到Visual Studio Code中的由LLM驱动的智能体，通过直接锚定到被分析代码的自主解释来支持程序理解。基于轻量级心理理论范式构建，TARS剖析开发者的专业知识、角色和风格偏好，然后相应地调整解释的深度和语气，并通过检索增强生成将其与项目文档联系起来。为评估TARS，我们对18名参与者就非平凡Java代码片段进行了对照实验。使用TARS的参与者完成任务速度快26%，报告的认知负荷更低，且发现解释能根据他们的个人资料进行有意义的调整。
英文摘要
Code comprehension is one of the most time-consuming tasks in software engineering, yet most LLM-based assistants produce explanations that ignore who is asking and force developers into a disruptive copy-paste workflow. We present TARS, an LLM-powered agent integrated into Visual Studio Code that supports program comprehension through autonomous explanations anchored directly to the code under analysis. Built around a lightweight Theory of Mind paradigm, TARS profiles a developer's expertise, role, and stylistic preferences, then adapts the depth and tone of its explanations accordingly, grounding them in project documentation via Retrieval-Augmented Generation. To evaluate TARS, we conducted a controlled experiment with 18 participants on non-trivial Java snippets. Participants using TARS completed tasks 26\% faster, reported lower cognitive load, and found the explanations meaningfully adapted to their profiles.
CommentsAccepted at ICSME 2026, Tool Demonstration and Data Showcase Track
