LLM Agents for Automated Dependency Upgrades

2025-10-03

AI Summary

Dependency library updates frequently introduce compatibility issues in software maintenance. Method: This paper proposes a large language model (LLM)-based multi-agent automated upgrade framework specifically designed for Java projects. The framework comprises three coordinated agents—summarization, control, and code agents—that jointly perform migration document parsing, fine-grained code context understanding, and change strategy generation to precisely locate dependency usage sites and apply safe, context-aware fixes. Contribution/Results: It is the first work to introduce a structured multi-agent architecture to dependency migration, effectively balancing semantic fidelity and engineering practicality. Evaluated on industrial-scale use cases, the framework reduces average token consumption and achieves a 71.4% upgrade accuracy—substantially outperforming existing approaches in both efficiency and reliability.

Technology Category

• Multiagent Systems: Teamwork
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

As a codebase expands over time, its library dependencies can become outdated and require updates to maintain innovation and security. However, updating a library can introduce breaking changes in the code, necessitating significant developer time for maintenance. To address this, we introduce a framework of LLM agents to be used in combination with migration documentation to automatically recommend and apply code updates and ensure compatibility with new versions. Our solution can automatically localize updated library usages in live Java codebases and implement recommended fixes in a user-friendly manner. The system architecture consists of multiple key components: a Summary Agent, Control Agent, and Code Agent. To validate our approach, we apply the framework on an industrial use case by which we create three synthetic code repositories with major Upgrade changes and benchmark our approach against state-of-the-art methods. Results show that our approach not only performs upgrades using fewer tokens across all cases but also achieves a precision of 71.4%, highlighting its efficiency and effectiveness compared to state-of-the-art methods.

Problem

• Automating library dependency upgrades to maintain code security
• Reducing developer maintenance time for dependency updates
• Ensuring code compatibility during automated dependency upgrades

Innovation

• LLM agents automate dependency upgrades using migration documentation
• Framework localizes library usage and implements compatibility fixes
• Multi-agent architecture reduces token usage while maintaining precision