授权闭包图：面向指令演化的LLM智能体的最小修复
Authorization Closure Graph: Minimal Repair for LLM Agents with Evolving User Instructions
AI总结：
针对LLM智能体指令部分变化时的授权更新问题，提出授权闭包图框架，选择性撤销受影响授权并计算最小修复，提升动作安全率和任务成功率。
AI中文摘要：
使用工具的大型语言模型（LLM）智能体越来越多地执行需要用户授权的状态改变操作。然而，当指令仅部分变化时，现有方法缺乏一种有原则的机制来选择性更新先前的授权。为此，我们提出了一种基于授权闭包图（Authorization-Closure-Graph, ACG）的框架，将授权及其依赖关系表示为演化的、带版本的状态。ACG选择性撤销受修订影响的授权，同时保留授权状态中未受影响的部分，并计算最小修复，仅识别执行所需的缺失证据或授权。这使得智能体能够适应修订后的指令，同时避免过时授权和不必要的授权请求。我们在两个自然任务中使用三种先进LLM评估了ACG，ACG持续提高了动作安全率和任务成功率。代码可在该URL获取。
英文摘要：
Tool-using large language model (LLM) agents increasingly perform state-changing actions that require user authorization. Yet existing approaches do not provide a principled mechanism for selectively updating prior authorization when only part of an instruction changes. To this end, we propose an Authorization-Closure-Graph (ACG)-based framework that represents authorization and its dependencies as an evolving, versioned state. ACG selectively invalidates authority affected by a revision while preserving unaffected portions of the authorization state, and computes a minimal repair that identifies only the missing evidence or authority required for execution. This enables agents to adapt to revised instructions while avoiding stale authority and unnecessary authorization requests. We evaluate ACG across three advanced LLMs in two natural tasks, and ACG consistently improves action safety rate and task success rate. Code is available at https://github.com/weiliang822/ACG.
