arXiv cs.CY· Sihao Liu, Ligeng Zhu, Zijian Zhang, Dongyun Zou, Zhengyang Zhang, Changye Li, Song Bian, Song Han, Tony Nowatzki·· 10 天前AI 评分62
Humanize：面向智能体编码的判断工程多智能体编排工作流
Humanize: Judgement Engineering for Agentic Coding
AI 导读
论文提出 Humanize，一种围绕判断工程构建的智能体编码多智能体编排工作流，在规划、实现、评审与学习之间设置显式且机械执行的决策边界。人类批准计划契约，builder 智能体分轮实现，来自另一厂商的 reviewer 智能体判定是否完成，由确定性 hook 而非模型在角色间路由工作并执行 72 道机械门。
正文 · 原文
Abstract:Agentic coding makes code generation cheap, but reliable completion remains difficult: the agent that writes the code is a weak judge of whether it is done.
We present Humanize, a multi-agent orchestration workflow for agentic coding built around judgement engineering: explicit, mechanically enforced decisions at the boundaries between planning, implementation, review, and learning. A human approves a plan contract, a builder agent implements it in rounds, and a reviewer agent from another vendor decides completion; deterministic hooks, not a model, route work between these roles and enforce 72 mechanical gates. Viewed as a Markov chain over repository states, alternating builder and reviewer samples jointly from two models, so a defect survives only if both miss it. We study Humanize through its deployment, 118 public postmortems of real loops, and its applications. Over 68 versions in 108 days, it gathered 1,468 GitHub stars.
Applications include a 567-file gem5 build-system migration under upstream review; Kernel Design Agents, which extend the loop with a kernel knowledge base and profiling feedback and placed in the top three of all three Full-Agent tracks of the MLSys 2026 FlashInfer contest; and, through Humanize Olympiad Agents (HOA), full scores in IOI 2026, IMO 2026, IPhO 2026, and IBO 2024, 418.5/437 in IChO 2026 (gold-medal). Humanize also achieves 672/672 on PutnamBench and ranks first (251/303) on Lean-Eval's leaderboard even competiting with professional mathematicians. The postmortems show that independent review catches unsupported builder claims, but stopping remains a key weakness. In reports that separate rounds by phase, two thirds of rounds occurred after implementation was accepted. This evidence is observational, not a controlled comparison of workflows.
| Subjects: | Artificial Intelligence (cs.AI); Computers and Society (cs.CY) |
| Cite as: | arXiv:2610.08900 [cs.AI] |
| (or arXiv:2610.08900v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.08900 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Sihao Liu [view email]
[v1]
Tue, 6 Oct 2026 17:54:42 UTC (266 KB)
来源：arXiv cs.CY · arxiv.org