Unverified50% confidenceFactExact time
Agent的失败-修正-再尝试过程被称为ReAct（Reasoning + Acting）模式或反思循环
1
Sources
50%
Confidence
Long-term
Relevance
7/6/2026
First Seen
Sources
Connecting an Open-Source AI Assistant to OpenClaw: A Complete Local Agent Setup Guide
bilibili枫影剑wind7/4/2026
Related Claims
UnverifiedThe "perception-decision-action" loop is the technical foundation of the ReAct (Reasoning + Acting) paradigm.76% similarVerifiedThe ReAct pattern's execution flow includes six steps: the user raises a request, the agent reasons (Reasoning), executes an action (Acting), observes the result (Observation), loops repeatedly, and returns an answer76% similarUnverifiedReAct的思考—行动—观察循环机制使Agent在处理多步骤复杂任务时表现出更高可靠性，显著减少幻觉和错误累积73% similarUnverifiedEarly agent products mostly adopted the ReAct (Reasoning + Acting) loop pattern, but this pattern has problems such as a single-step failure causing a global collapse and the context growing larger with each step.72% similarUnverified错误自动修复能力在技术上依托反思型Agent（Reflexion Agent）机制，编译器返回的错误日志被Agent捕获后作为新输入触发下一轮推理与修正72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/113674API
curl https://kongchang.com/api/v1/knowledge/claims/113674MCP
get_claim(id=113674)