SWE-AGILE: A Software Agent Framework for Efficiently Managing Dynamic Reasoning Context

2026-04-13

AI Summary

Existing autonomous software engineering approaches often struggle with complex edge cases due to insufficient systematic reasoning, leading to context explosion or repetitive inference. This work proposes a dynamic reasoning context management framework that, for the first time, integrates a sliding window mechanism with reasoning summary compression into autonomous software engineering. By combining an extended chain-of-thought (CoT) strategy, the framework preserves recent detailed reasoning while compressing historical information, effectively balancing context redundancy and information loss across multi-turn tasks. Evaluated on SWE-Bench-Verified using only 2.2k trajectories and 896 tasks, the method establishes a new state-of-the-art performance for models in the 7B–8B parameter range.

Technology Category

• Knowledge Representation and Reasoning: Automated Reasoning and Theorem Proving
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Search and Optimization: Metareasoning and Metaheuristics

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Prior representative ReAct-style approaches in autonomous Software Engineering (SWE) typically lack the explicit System-2 reasoning required for deep analysis and handling complex edge cases. While recent reasoning models demonstrate the potential of extended Chain-of-Thought (CoT), applying them to the multi-turn SWE task creates a fundamental dilemma: retaining full reasoning history leads to context explosion and ``Lost-in-the-Middle'' degradation, while discarding it would force the agent to redundantly re-reason at every step. To address these challenges, we propose SWE-AGILE, a novel software agent framework designed to bridge the gap between reasoning depth, efficiency, and context constraints. SWE-AGILE introduces a Dynamic Reasoning Context strategy, maintaining a ``sliding window'' of detailed reasoning for immediate continuity to prevent redundant re-analyzing, while compressing historical reasoning content into concise Reasoning Digests. Empirically, SWE-AGILE sets a new standard for 7B-8B models on SWE-Bench-Verified using only 2.2k trajectories and 896 tasks. Code is available at https://github.com/KDEGroup/SWE-AGILE.

Problem

• Software Engineering
• Reasoning Context
• Chain-of-Thought
• Autonomous Agents
• Context Management

Innovation

• Dynamic Reasoning Context
• Reasoning Digests
• Sliding Window
• SWE Agent
• Context Compression