跳到正文
arXiv:cs.AI· Bowen Xu, Boyu Chen·· 3 小时前

跨提供商审查作为编码智能体的运行时契约:一项受控试点与故障注入研究

Cross-Provider Review as a Runtime Contract for Coding Agents: A Controlled Pilot and Fault-Injection Study

AI 导读

研究提出一种跨提供商审查契约,让不同资源池的第二个编码智能体审查已完成答案,具备有界执行、受限审查能力与显式失败状态。20 组配对开发轮次中 8 次出现实质性审查发现(95% 精确区间 19.1%-63.9%)。边界条件扫描复现了部分输入下的假成功(四个截断级别此前通过、修复后失败),并修复了进程回收期间的取消问题;CLI 探测中 Codex 在五次只读试验中五次尝试写入均失败。

正文

View PDF HTML (experimental)

Abstract:Coding agents increasingly share a workstation while drawing on separate providers and subscription allowances. A second agent can inspect a completed answer, but the call spends another pool and may provide no substantive finding. We describe an advisory cross-provider review contract: distinct resource pools, bounded execution, restricted reviewer capabilities, complete input delivery, usable semantic output, explicit failure states and durable per-attempt evidence. In a controlled, agent-authored pilot of 20 paired development turns, eight had a material reviewer finding (95% exact interval 19.1-63.9%). A boundary-condition scan across both reviewer backends reproduced a previously discovered false success on partial input: four truncation levels passed historically and failed after repair. The scan also found and repaired cancellation during process reaping. In real CLI probes, Claude had no writing tools; Codex attempted writes in five of five read-only trials, each write tool failed, and no disposable repository changed. These tests cover specified paths and versions, not field reliability. A preregistered shadow study of metadata-only review allocation accrued 25 formal observations before an exact-runtime regression found a third defect: a reviewer exiting nonzero with a well-formed verdict was counted as complete. Exit status was not recorded per attempt, so exposure cannot be resolved retrospectively. The 25 formal and two pending records remain an audit cohort; the measurement-valid cohort restarted at zero and collection has begun. No gate result is reported.
Comments: 11 pages, 1 figure, 3 tables. Project page: this https URL
Subjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI)
Cite as: arXiv:2610.10961 [cs.SE]
  (or arXiv:2610.10961v1 [cs.SE] for this version)
  https://doi.org/10.48550/arXiv.2610.10961

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Bowen Xu [view email]
[v1] Wed, 7 Oct 2026 22:28:22 UTC (52 KB)

来源:arXiv:cs.AI · arxiv.org