无Docker：面向编码代理的无环境程序验证器

摘要

程序验证器在训练代码智能体（coding agents）中扮演核心角色，包括为监督微调（SFT）筛选轨迹、为强化学习（RL）提供奖励。标准基于执行的验证需要运行单元测试（位于每个仓库环境内，如Docker镜像），这会产生高昂的环境搭建成本。我们提出Dockerless——一种无环境的智能体式补丁验证器，无需执行代码即可评估生成的代码补丁。Dockerless并非简单地将候选补丁与参考方案匹配，而是通过智能仓库探索收集证据来判断补丁正确性。在验证器评估基准上，Dockerless比最强的开源验证器高出14.3个AUC点。将Dockerless同时用作SFT轨迹筛选器和RL奖励，可实现完全无环境的训练后流程。由此得到的模型在SWE-bench Verified、Multilingual和Pro上分别达到62.0%、50.0%和35.2%的解决率，相比Qwen3.5-9B基线分别提升2.4、8.7和2.9个百分点，与基于环境的训练后流程性能相当。

English

Program verifiers play a central role in training coding agents, including selecting trajectories for supervised fine-tuning (SFT) and providing rewards for reinforcement learning (RL). Standard execution-based verification requires running unit tests inside per-repository environments such as Docker images, incurring substantial environment setup costs. We propose Dockerless, an environment-free agentic patch verifier that evaluates generated code patches without executing them. Rather than simply matching candidate patches to references, Dockerless judges patch correctness using evidence gathered through agentic repository exploration. On a verifier evaluation benchmark, Dockerless outperforms the strongest open-source verifier by 14.3 AUC points. Using Dockerless as both the SFT trajectory filter and the RL reward enables a fully environment-free post-training pipeline. The resulting model reaches 62.0%, 50.0%, and 35.2% resolve rate on SWE-bench Verified, Multilingual, and Pro, respectively. It surpasses the Qwen3.5-9B baseline by 2.4, 8.7, and 2.9 points, matching environment-based post-training.