dorsal/arxiv
View SchemaMemGovern: Enhancing Code Agents through Learning from Governed Human Experiences
| Authors | Qihao Wang, Ziming Cheng, Shuo Zhang, Fan Liu, Rui Xu, Heng Lian, Kunyi Wang, Xiaoming Yu, Jianghao Yin, Sen Hu, Yue Hu, Shaolei Zhang, Yanbing Liu, Ronghao Chen, Huacan Wang |
|---|---|
| Categories | |
| ArXiv ID | 2601.06789vv1 |
| URL | https://arxiv.org/abs/2601.06789 |
| License | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ |
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.
{
"annotation_id": "b76cd726-99ec-41d0-a14b-4c63ff443c97",
"date_created": "2026-02-17T05:53:08.816000Z",
"date_modified": "2026-02-17T05:53:08.816000Z",
"file_hash": "e898677102b43613600cc7df2a9139e00cb99374f1f54f1ca47da887d336ff13",
"private": false,
"record": {
"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.",
"arxiv_id": "2601.06789",
"authors": [
"Qihao Wang",
"Ziming Cheng",
"Shuo Zhang",
"Fan Liu",
"Rui Xu",
"Heng Lian",
"Kunyi Wang",
"Xiaoming Yu",
"Jianghao Yin",
"Sen Hu",
"Yue Hu",
"Shaolei Zhang",
"Yanbing Liu",
"Ronghao Chen",
"Huacan Wang"
],
"categories": [
"cs.SE",
"cs.AI"
],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"title": "MemGovern: Enhancing Code Agents through Learning from Governed Human Experiences",
"url": "https://arxiv.org/abs/2601.06789",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "80062de5-7e1b-46b1-849e-f22ff556a230",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}