Local Edits, Global Ripples: Replay-Informed Policy Adaptation for Workflow Synthesis
Source
Manqing Mao, Hong Wang, Samson Koelle, Jie Yuan, Zhuoer Wang, James Feng, Yanjun Lin, Daniel Edmiston, Nikki Lijing Kuang, Zhecheng Sheng, Wei Niu
Author
Manqing Mao, Hong Wang, Samson Koelle, Jie Yuan, Zhuoer Wang, James Feng, Yanjun Lin, Daniel Edmiston, Nikki Lijing Kuang, Zhecheng Sheng, Wei Niu
Date
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
Why it matters
Gives a concrete technique for safely accumulating prompt-level fixes to workflow-generating agents over time without retraining, addressing a real failure mode where one fix silently breaks or interacts badly with another.
articleAgentR A Stateful and Recovery-Aware Software Architecture for LLM-based Auditable WorkflowsRiya Samanta, Bidyut Saha, Soumya Kanti Ghosh, Rajkumar Buyya
articleOpenAgentFlow: Enabling System-Wide Safety Boundaries for Heterogeneous AI Agent FleetsDongsheng Chen, Xiangyu Zhao, Xin Yao, Xuetao Wei
articleBeyond Prompt Engineering: A Systematic Analysis of Prompt Lexical Sensitivity and Its Impacts on QualityQipeng Xie, Zi Liang, Jiafei Wu, Yufei Chen, Weizheng Wang, Wenao Ma, Zhong Ming, Haiqin Yang, Kaishun Wu