Reproduce “Rethinking Tool Design for Agentic RCA: A Controlled Empirical Study” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yu Luo, Rongchen Gao, Zhenhui Zhou, Changchang Liu, Yuliang You, Yongqian Sun, Shenglin Zhang, Qiuai Fu and 2 more
The study compares tool configurations for metric-based microservice diagnosis, separating metric access, evidence analysis, and diagnostic recommendations. Across 375 cases, Qwen3.7-Plus reaches 82.8% AC@1 with L3 alone, versus 85.4% with a general Python interface, while using fewer tokens and less elapsed time. A full toolset improves fault-type accuracy but lowers localization accuracy in this tested setting.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.