Unverified50% confidenceFactExact time
The knowledge base keeps a RAW directory for storing the most original notes or web page information, from which the agent embeds and distills knowledge and automatically files it into the corresponding directories, separating raw information from structured knowledge.
1
Sources
50%
Confidence
Long-term
Relevance
10/28/2026
First Seen
Sources
From Bookmarks to Second Brain: Auto-Building a Knowledge Base with Feishu + AI
bilibili不吃辣的Chris10/28/2026
Related Entities
Related Claims
UnverifiedPersistent knowledge has the core characteristics of being versionable, reviewable, and traceable, and is typically stored in a Git repository or vector database in the form of documents or structured data.72% similarUnverifiedDuring operation, agents leave discrete data traces (such as conversation history, tool call records, retrieved document snippets and intermediate reasoning content), and these traces are distributed across different storage layers such as application logs, vector databases and model provider caches.70% similarUnverifiedIn complex scenarios, knowledge bases typically use vector databases (such as Pinecone, Weaviate, and Qdrant) to store embedding-processed documents for semantic retrieval.68% 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68% similarUnverifiedThe system's knowledge base documents are stored in a Redis vector database after being processed through chunking.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1107669API
curl https://kongchang.com/api/v1/knowledge/claims/1107669MCP
get_claim(id=1107669)