Beyond Textual Repository Exploration: Dual-Modal Structural Reasoning for Agentic Issue Resolution

2026-07-02

AI Summary

Existing code agents rely solely on textual navigation, struggling with long-range dependencies in large-scale codebases and often suffering from exploration drift and missed localization. This work proposes DUALVIEW, a novel framework that introduces, for the first time, a dual-modality structured representation combining textual and visual modalities to guide repository-level agent exploration. DUALVIEW constructs four complementary graph views—Module Coupling Graph (MCG), Function Call Graph (FCG), Class Hierarchy Graph (CHG), and Program Dependency Graph (PDG)—and provides queryable interfaces to support persistent dependency reasoning. Evaluated on the SWE-bench Pro and Verified benchmarks, DUALVIEW significantly improves problem-resolution success rates across multiple agent architectures. Ablation studies confirm that externalized visual structures play a critical role in enabling effective long-range exploration.

Abstract

Recent advances in agentic program repair have significantly improved issue resolution by enabling iterative repository exploration. However, existing approaches predominantly rely on sequential, text-based code navigation, which fundamentally limits their ability to reason over large-scale long-horizon repositories with complex and long-range dependencies. As issue-resolution agents traverse repositories through fragmented textual observations, structural information such as module organization, call relationships, and dependency chains must be repeatedly reconstructed across interaction steps, often leading to exploration drift and incomplete localization. We present DUALVIEW, a dual-modal structural scaffolding framework that brings visual reasoning into repository exploration for issue-resolution agents. DUALVIEW represents repository structure through four complementary graph views: Module Coupling Graph (MCG), Function Call Graph (FCG), Class Hierarchy Graph (CHG), and Program Dependence Graph (PDG), and exposes them through a queryable interface with visual and textual responses. Rather than reconstructing repository structure from a sequence of textual observations, agents can directly reason over persistent visual representations of code dependencies, enabling more effective exploration and understanding of long-horizon codebases. We evaluate DUALVIEW on SWE-bench Pro and Verified. Results show that DUALVIEW consistently improves issue-resolution performance across different agent architectures and model families. Further ablation studies demonstrate that the gains arise not only from textual structural information but also from visual externalization of repository dependencies, which better supports long-horizon repository exploration.

Problem

• repository exploration
• structural reasoning
• long-horizon dependencies
• issue resolution
• code navigation

Innovation

• dual-modal reasoning
• structural scaffolding
• visual code representation
• repository exploration
• agentic program repair