Saved in:
Bibliographic Details
Main Authors: Wang, Qihao, Cheng, Ziming, Zhang, Shuo, Liu, Fan, Xu, Rui, Lian, Heng, Wang, Kunyi, Yu, Xiaoming, Yin, Jianghao, Hu, Sen, Hu, Yue, Zhang, Shaolei, Liu, Yanbing, Chen, Ronghao, Wang, Huacan
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.06789
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914250439524352
author Wang, Qihao
Cheng, Ziming
Zhang, Shuo
Liu, Fan
Xu, Rui
Lian, Heng
Wang, Kunyi
Yu, Xiaoming
Yin, Jianghao
Hu, Sen
Hu, Yue
Zhang, Shaolei
Liu, Yanbing
Chen, Ronghao
Wang, Huacan
author_facet Wang, Qihao
Cheng, Ziming
Zhang, Shuo
Liu, Fan
Xu, Rui
Lian, Heng
Wang, Kunyi
Yu, Xiaoming
Yin, Jianghao
Hu, Sen
Hu, Yue
Zhang, Shaolei
Liu, Yanbing
Chen, Ronghao
Wang, Huacan
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2601_06789
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle MemGovern: Enhancing Code Agents through Learning from Governed Human Experiences
Wang, Qihao
Cheng, Ziming
Zhang, Shuo
Liu, Fan
Xu, Rui
Lian, Heng
Wang, Kunyi
Yu, Xiaoming
Yin, Jianghao
Hu, Sen
Hu, Yue
Zhang, Shaolei
Liu, Yanbing
Chen, Ronghao
Wang, Huacan
Software Engineering
Artificial Intelligence
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.
title MemGovern: Enhancing Code Agents through Learning from Governed Human Experiences
topic Software Engineering
Artificial Intelligence
url https://arxiv.org/abs/2601.06789