用智能代理自动构建可靠代码环境，提升开发智能体的训练效率。
DockSmith: Scaling Reliable Coding Environments via an Agentic Docker Builder
- 设计智能代理自动处理依赖、错误恢复，实现长时序环境构建。
- 在 Multi-Docker-Eval 上达 39.72% 失败转成功率，58.28% 提交率。
- 不仅优化构建过程，还增强智能体在多任务上的泛化能力。
基于 Docker 的环境构建是规模化执行驱动型软件工程智能体训练与评估的主要瓶颈。本文提出 DockSmith，一种专用的智能体式 Docker 构建工具。它将环境构建视为核心智能体能力，涵盖长周期工具使用、依赖推理和故障恢复，生成可迁移的监督信号。DockSmith 在由 SWE-Factory 式流水线（含循环检测控制器和跨任务成功记忆）生成的大规模、执行驱动型构建轨迹上训练，使用 30B-A3B 模型，在 Multi-Docker-Eval 上达到 39.72% 的失败转成功率和 58.28% 的提交率，为开源模型当前最优表现。此外，其在 SWE-bench Verified、SWE-bench Multilingual 与 Terminal-Bench 2.0 上均展现更强的分布外性能，证明环境构建本身具备更广泛的智能体增益。
原文摘要 · Abstract (English)
Reliable Docker-based environment construction is a dominant bottleneck for scaling execution-grounded training and evaluation of software engineering agents. We introduce DockSmith, a specialized agentic Docker builder designed to address this challenge. DockSmith treats environment construction not only as a preprocessing step, but as a core agentic capability that exercises long-horizon tool use, dependency reasoning, and failure recovery, yielding supervision that transfers beyond Docker building itself. DockSmith is trained on large-scale, execution-grounded Docker-building trajectories produced by a SWE-Factory-style pipeline augmented with a loop-detection controller and a cross-task success memory. Training a 30B-A3B model on these trajectories achieves open-source state-of-the-art performance on Multi-Docker-Eval, with 39.72% Fail-to-Pass and 58.28% Commit Rate. Moreover, DockSmith improves out-of-distribution performance on SWE-bench Verified, SWE-bench Multilingual, and Terminal-Bench 2.0, demonstrating broader agentic benefits of environment construction.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。