Knowledge Vault (Powered by TiDB Zero)

Overview

7w4.net提供免費和付費技能下載。

Knowledge Vault is a Long-Term Memory module for AI Agents, powered by TiDB Vector Search (RAG).

Traditional agent memory (context window) is ephemeral and limited. Knowledge Vault allows agents to:

1. Store: Ingest documents, notes, and facts as vector embeddings.
2. Retrieve: Semantically search for relevant information based on user queries ("RAG").
3. Remember: Access unlimited historical context without overflowing the LLM prompt.

Why use this?

• Infinite Recall: Store millions of documents without confusing the agent.
• Contextual Relevance: Find exact paragraphs related to a question, not just keywords.
• Privacy: Keep your knowledge base private in your own TiDB Cloud instance.

Prerequisites

• TiDB Cloud (Serverless): With Vector Search enabled.
• Embedding Model: Requires GEMINI_API_KEY (or compatible).

🔐 Security & Provisioning

This skill operates in two modes:

1. Bring Your Own Database (Recommended): Set TIDB_HOST, TIDB_USER, TIDB_PASSWORD environment variables. The skill will use your existing database.
2. Auto-Provisioning (Fallback): If no credentials are found, the skill calls the TiDB Zero API to create a temporary, ephemeral database for you. It caches the connection string locally (~/.openclaw_knowledge_vault_dsn) to persist memory across runs.

Installation

1. Add to TOOLS.md

- **knowledge-vault**: Store and retrieve knowledge using vector search.
  - **Location:** `{baseDir}/skills/knowledge_vault/SKILL.md`
  - **Command:** `python {baseDir}/skills/knowledge_vault/run.py --action search --query "<QUESTION>"`

2. Add to AGENTS.md (Protocol)

Copy PROTOCOL.md.

Usage

• Add Knowledge:

  python {baseDir}/run.py --action add --content "The user prefers spicy food but is allergic to peanuts."

• Search (RAG):

  python {baseDir}/run.py --action search --query "What are the user's dietary restrictions?"

這個 Skill 整體質量不錯，能較好地實現長期記憶儲存和語義搜尋功能。文件清晰、配置靈活是它的亮點。但存在一些小問題：嵌入維度設定可能與實際模型不符，錯誤處理不夠完善，而且缺少測試保證穩定性。對於需要穩定使用的場景，建議先確認維度配置與實際環境匹配後再部署。