Saved in:
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2601.22859 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915768327733248 |
|---|---|
| author | Guo, Chuanzhe Wu, Jingjing He, Sijun Chen, Yang Kuang, Zhaoqi Fan, Shilong Chen, Bingjin Bao, Siqi Liu, Jing Wu, Hua Zhu, Qingfu Che, Wanxiang Wang, Haifeng |
| author_facet | Guo, Chuanzhe Wu, Jingjing He, Sijun Chen, Yang Kuang, Zhaoqi Fan, Shilong Chen, Bingjin Bao, Siqi Liu, Jing Wu, Hua Zhu, Qingfu Che, Wanxiang Wang, Haifeng |
| contents | The evolution of Large Language Model (LLM) agents for software engineering (SWE) is constrained by the scarcity of verifiable datasets, a bottleneck stemming from the complexity of constructing executable environments across diverse languages. To address this, we introduce MEnvAgent, a Multi-language framework for automated Environment construction that facilitates scalable generation of verifiable task instances. MEnvAgent employs a multi-agent Planning-Execution-Verification architecture to autonomously resolve construction failures and integrates a novel Environment Reuse Mechanism that reduces computational overhead by incrementally patching historical environments. Evaluations on MEnvBench, a new benchmark comprising 1,000 tasks across 10 languages, demonstrate that MEnvAgent outperforms baselines, improving Fail-to-Pass (F2P) rates by 8.6% while reducing time costs by 43%. Additionally, we demonstrate the utility of MEnvAgent by constructing MEnvData-SWE, the largest open-source polyglot dataset of realistic verifiable Docker environments to date, alongside solution trajectories that enable consistent performance gains on SWE tasks across a wide range of models. Our code, benchmark, and dataset are available at https://github.com/ernie-research/MEnvAgent. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_22859 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MEnvAgent: Scalable Polyglot Environment Construction for Verifiable Software Engineering Guo, Chuanzhe Wu, Jingjing He, Sijun Chen, Yang Kuang, Zhaoqi Fan, Shilong Chen, Bingjin Bao, Siqi Liu, Jing Wu, Hua Zhu, Qingfu Che, Wanxiang Wang, Haifeng Software Engineering Artificial Intelligence The evolution of Large Language Model (LLM) agents for software engineering (SWE) is constrained by the scarcity of verifiable datasets, a bottleneck stemming from the complexity of constructing executable environments across diverse languages. To address this, we introduce MEnvAgent, a Multi-language framework for automated Environment construction that facilitates scalable generation of verifiable task instances. MEnvAgent employs a multi-agent Planning-Execution-Verification architecture to autonomously resolve construction failures and integrates a novel Environment Reuse Mechanism that reduces computational overhead by incrementally patching historical environments. Evaluations on MEnvBench, a new benchmark comprising 1,000 tasks across 10 languages, demonstrate that MEnvAgent outperforms baselines, improving Fail-to-Pass (F2P) rates by 8.6% while reducing time costs by 43%. Additionally, we demonstrate the utility of MEnvAgent by constructing MEnvData-SWE, the largest open-source polyglot dataset of realistic verifiable Docker environments to date, alongside solution trajectories that enable consistent performance gains on SWE tasks across a wide range of models. Our code, benchmark, and dataset are available at https://github.com/ernie-research/MEnvAgent. |
| title | MEnvAgent: Scalable Polyglot Environment Construction for Verifiable Software Engineering |
| topic | Software Engineering Artificial Intelligence |
| url | https://arxiv.org/abs/2601.22859 |