这篇论文告诉你：给编码智能体写AGENTS.md时，别一次性写好就完，得用探针-改进法迭代修。实测在SWE-bench上解决率涨了近5个点，主要是能让智能体多搞定14.5%的实例。
论文提出探针-改进调优法，通过合成bug修复探针迭代诊断和修补仓库指导文件，无需代理循环或工具调用。在SWE-bench Verified上，使用Qwen3.5-35B-A3B模型200步，平均解决率33.0%，高于静态知识库的28.3%和无指导基线的25.5%（p<0.001）。改进来自覆盖度而非精确度：优化指导使可评估补丁增加14.5个百分点，但补丁精确度稳定在约59%（p=0.119）。跨模型实验显示，当模型无法生成足够诊断输出时调优效果下降，但补丁精确度仍保持恒定。
Probe-and-Refine Tuning of Repository Guidance for Coding Agents
LLM-based coding agents need higher-level operational knowledge about a repository (which files house which subsystems, how to run the test suite, which workflows have historically led to wrong fixes) that does not exist in the code itself. Engineers typically maintain \texttt{AGENTS.md} files to supply this context as instructions for coding agents, but whether they help is contested: recent studies disagree on whether LLM-generated guidance improves or harms agent performance. In this paper we show that how the guidance is produced is the decisive variable, and introduce \emph{probe-and-refine tuning}: a procedure that uses synthetic bug-fix probes to iteratively diagnose and patch a repository's guidance file through single-shot LLM calls, with no agent loop or tool use during tuning. On SWE-bench Verified across four independent trials with Qwen3.5-35B-A3B at 200 steps, probe-and-refine achieves 33.0\,\% mean resolve rate vs.\ 28.3\,\% for the static knowledge base used to initialize it and 25.5\,\% for an unguided baseline ($p < 0.001$ for both probe-and-refine contrasts). The improvement comes from coverage rather than precision: refined guidance produces evaluable patches for 14.5 percentage points (pp) more instances while per-patch precision remains statistically constant ($\sim$59\,\%, $p = 0.119$), showing that improved guidance helps agents reach the correct file rather than improving the quality of the changes they make. Further, a step-budget experiment shows that guidance is what lets the agent use a larger step budget productively, and a cross-model experiment with NVIDIA-Nemotron-3-Nano-30B-A3B finds that the tuning loop degrades when the model cannot generate sufficiently diagnostic output, though per-patch precision remains constant even then.