TARS: A Theory-of-Mind Agent for Personalized In-IDE Code Comprehension

2026-07-17

AI Summary

This work addresses a critical limitation in existing code comprehension tools, which often overlook individual developer differences and rely on disruptive copy-paste interactions that interrupt workflow. To overcome these issues, the authors propose TARS, an intelligent agent integrated into VS Code that introduces lightweight Theory of Mind into the IDE for the first time. TARS models developers’ expertise, roles, and stylistic preferences and leverages retrieval-augmented generation (RAG) to deliver personalized code explanations grounded in the immediate code context. User studies demonstrate that TARS improves task completion speed by 26%, significantly reduces cognitive load, and generates explanations that align more closely with individual developer profiles.

Abstract

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.

Problem

• code comprehension
• personalization
• Theory of Mind
• developer profiling
• in-IDE assistance

Innovation

• Theory of Mind
• Personalized Code Comprehension
• Retrieval-Augmented Generation
• In-IDE Agent
• LLM-powered Assistant