ASAD: Adaptive Software Agents for Debugging

2026-09-30

AI Summary

This study addresses the architectural rigidity of existing multi-agent debugging frameworks, which struggle to accommodate heterogeneous defects and consequently suffer from resource misallocation. To overcome this limitation, we propose a large language model-based adaptive multi-agent debugging system that introduces a novel defect-analysis-driven dynamic team assembly mechanism. Through iterative planning-reflection-execution and dynamic orchestration techniques, the system configures the number of agents, their roles, and collaboration strategies in real time according to defect complexity. Experimental results across multiple benchmarks demonstrate that the proposed approach improves repair rates by 12%–20% and outperforms static systems in accuracy by 4%–9%, while reducing average agent invocations by 32%. These findings confirm that the method achieves precise and efficient resource allocation for automated software debugging.

Abstract

The integration of Large Language Models (LLMs) into multi-agent systems has shown great potential for automated debugging. Yet nearly all current frameworks rely on rigid, predefined architectures: the number of agents, their roles, and their interaction patterns are fixed before any analysis of the bug occurs. This one-size-fits-all approach is fundamentally mismatched to the heterogeneous nature of software defects. Simple bugs waste resources on unnecessary coordination, while complex ones suffer from insufficient or poorly aligned expertise. This paper introduces ASAD, an adaptive agentic system for debugging that configures its team according to the nature and complexity of each bug. ASAD initiates the debugging process by analyzing the faulty code and dynamically determines the number of agents to deploy, the specialized roles they should have, and the collaboration strategy they should follow. A central coordinator orchestrates this process through iterative planning, reflection, and execution; applying fast single-pass repairs for simple issues while assembling purpose-built teams to tackle more complex failures. We evaluate ASAD on three established benchmarks: Defects4J, DebugBench, and CodeFlaws, using multiple LLMs, such as DeepSeek-V3, Qwen-3 and GPT-5. ASAD consistently improves bug-fix rates by 12--20% over chain-of-thought(CoT) prompting and consistently outperforms static multi-agent systems by 4--9% in fix precision while reducing average agent usage by 32%. Crucially, our system dynamically adjusts the number and roles of agents: it resolves simple bugs with minimal coordination and scales agent involvement only for more complex cases.

Problem

• automated debugging
• multi-agent systems
• large language models
• software defects
• adaptive agents

Innovation

• Adaptive Multi-Agent System
• Automated Debugging
• Dynamic Role Allocation
• Large Language Models
• Iterative Planning