MemGovern: Enhancing Code Agents through Learning from Governed Human Experiences

TL;DR

MemGovern is a framework that transforms raw GitHub data into structured experience cards, enabling code agents to leverage historical human expertise and significantly improve bug resolution performance.

Contribution

It introduces a novel experience governance framework and experience search strategy to enhance code agents with open-world human knowledge from GitHub.

Findings

Produced 135K governed experience cards.

Achieved 4.65% improvement in bug resolution rates.

Provides a plug-in memory infrastructure for code agents.

Abstract

While autonomous software engineering (SWE) agents are reshaping programming paradigms, they currently suffer from a "closed-world" limitation: they attempt to fix bugs from scratch or solely using local context, ignoring the immense historical human experience available on platforms like GitHub. Accessing this open-world experience is hindered by the unstructured and fragmented nature of real-world issue-tracking data. In this paper, we introduce MemGovern, a framework designed to govern and transform raw GitHub data into actionable experiential memory for agents. MemGovern employs experience governance to convert human experience into agent-friendly experience cards and introduces an agentic experience search strategy that enables logic-driven retrieval of human expertise. By producing 135K governed experience cards, MemGovern achieves a significant performance boost, improving resolution rates on the SWE-bench Verified by 4.65%. As a plug-in approach, MemGovern provides a solution for agent-friendly memory infrastructure.