Unverified50% confidenceDecision RuleExact time
可迁移方法论：识别领域实体与关系构建知识图谱、以工具形式暴露给LLM、通过多轮工具调用实现深层信息挖掘
1
Sources
50%
Confidence
Long-term
Relevance
7/17/2026
First Seen
Sources
Related Claims
Unverified在真实上下文工程场景中，上下文散布在多个来源：本地文件系统、数据库、Web、长期记忆等，每种上下文源需要对应的搜索工具62% similarUnverifiedSkill 中的 references 目录可注入领域知识文档，类似 RAG 的本地知识库，让 Agent 回答有据可查60% similarUnverifiedGBrain distills documents into entities and relationships through Named Entity Recognition (NER) and Relation Extraction (RE) techniques to form a knowledge network.60% similarUnverified编译式RAG（LLM Wiki）将非结构化文档经过内容解析、知识点提取、关系建模和冲突消解等步骤，转化为结构化的可导航知识网络59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/541855API
curl https://kongchang.com/api/v1/knowledge/claims/541855MCP
get_claim(id=541855)