推测性分叉：将空闲等待转为并行探索的仓库环境搭建
原标题：Speculative Forking: Turning Idle Wait into Parallel Exploration for Repository Environment Setup
AI 摘要
研究者提出 SpecBuild，一个基于 LLM 的仓库环境搭建系统，通过“推测性分叉”把原本串行的试错流程改为并行探索：策略级分叉同时尝试不同基础镜像与依赖策略，命令级分叉在慢命令失败前预先准备修复分支，谁先成功就采用谁。在 Multi-Docker-Eval（334 个任务、40 个仓库、9 种语言）上，三种骨干模型下成功率比最强基线提升 6.4 至 12.0 个百分点，平均搭建时间降低 22.4% 至 41.6%。
正文节选
Speculative Forking: Turning Idle Wait into Parallel Exploration for Repository Environment Setup Abstract Before developers can use a software project, they must first get it to run. This often requires installing the right language runtime, system packages, build tools, and test commands. Recent LLM-based agents automate environment setup by inspecting a repository, writing setup scripts, running them, and repairing the resulting errors. However, these agents still follow a sequential trial-an