mlsanityData to 7 Oct 2026

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

arXiv 2607.15948 ↗First posted 17 Jul 2026cs.SEcited by 0

On the map: LLM agents and tool use → Software engineering coding agents

Abstract (arXiv)

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.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use. Builds on 8 of its 27 references that are in the map. First posted on arXiv 17 Jul 2026.

Builds on

8 of 27 references.

References · 8

  1. 2021CodeXGLUE: A Machine Learning Benchmark Dataset for Code Understanding and Generationcited by 732
  2. 2019CodeSearchNet Challenge: Evaluating the State of Semantic Code Searchcited by 638
  3. 2023Using an LLM to Help With Code Understandingcited by 222
  4. 2023Think Twice: Perspective-Taking Improves Large Language Models' Theory-of-Mind Capabilitiescited by 86
  5. 2023Boosting Theory-of-Mind Performance in Large Language Models via Promptingcited by 64
  6. 2024Ivie: Lightweight Anchored Explanations of Just-Generated Codecited by 14
  7. 2024What You Need is What You Get: Theory of Mind for an LLM-Based Code Understanding Assistantcited by 8
  8. 2025Do Prompt Patterns Affect Code Quality? A First Empirical Assessment of ChatGPT-Generated Codecited by 4

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024From Critique to Clarity: A Pathway to Faithful and Personalized Code Explanations with Large Language Models
  2. 2026Vibe-GUIDE: A Graph-based User Interface in IDEs for Oversight in Vibe Coding
  3. 2026Do Personalized Skills Help Coding Agents? An Empirical Study of Developer Interaction Histories
  4. 2025TOM-SWE: User Mental Modeling For Software Engineering Agents
  5. 2026Beyond the 'Diff': Addressing Agentic Entropy in Agentic Software Development
  6. 2026CodeTaste: Can LLMs Generate Human-Level Code Refactorings?
  7. 2025Beyond Autocomplete: Designing CopilotLens Towards Transparent and Explainable AI Coding Agents
  8. 2025Wired for Reuse: Automating Context-Aware Code Adaptation in IDEs via LLM-Based Agent