Multi-agent LLM architectures offer opportunities for pedagogical agents to help students construct domain knowledge and develop critical-thinking skills, yet many operate on a "one-size-fits-all" basis, limiting their ability to provide personalized support. To address this, we introduce Evidence-Decision-Feedback (EDF), a theoretical framework for adaptive scaffolding using LLMs. EDF integrates elements of intelligent tutoring systems and agentic behavior by organizing interactions around evidentiary inference, pedagogical decision-making, and adaptive feedback. We instantiate EDF through Copa, an agentic collaborative peer agent for STEM+C problem-solving. In an authentic high school classroom study, we show that EDF-aligned interactions align feedback with students' demonstrated understanding and task mastery; promote gradual scaffold fading; and support interpretable, evidence-grounded explanations without fostering overreliance.

翻译：多智能体LLM架构为教学智能体提供了帮助学生构建领域知识和发展批判性思维技能的机会，然而许多系统采用"一刀切"模式，限制了其提供个性化支持的能力。为此，我们提出证据-决策-反馈（EDF）——一个基于LLM的自适应支架理论框架。EDF通过围绕证据推理、教学决策和自适应反馈组织交互，整合了智能导学系统与智能体行为的关键要素。我们通过Copa（一个面向STEM+C问题解决的协作式同伴智能体）实现了EDF框架。在真实的高中课堂研究中，我们证明EDF框架下的交互能够：使反馈与学生的显性理解及任务掌握程度相匹配；促进支架的渐进式消退；支持可解释、基于证据的说明，同时避免助长学生的过度依赖。