rag

Retrieve documents from a knowledge base and generate LLM answers.

Updated Aug 23, 2026

Search Tags:#semantic search #llm #rag #knowledge base #retrieval-augmented generation #document qa

Ready to run

npx skills add https://github.com/liushuang393/serverlessAIAgents --skill rag-liushuang393

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: rag
Source: https://github.com/liushuang393/serverlessAIAgents/tree/main/agentflow/skills/builtin/rag
Command: npx skills add https://github.com/liushuang393/serverlessAIAgents --skill rag-liushuang393

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The RAG Skill grounds AI-generated responses in an external knowledge base by retrieving relevant documents and generating context-aware answers with an LLM, reducing hallucinations and speeding up information discovery.

Core Features & Use Cases

Document ingestion and semantic search to fetch relevant passages
Contextual prompting and answer generation with retrieved context
Source tracking and topic management for auditable results
Use cases include internal knowledge bases, product documentation QA, policy lookup, and customer support knowledge bases.

Quick Start

Initialize the RAGSkill: rag = RAGSkill(); await rag.start()
Add documents: await rag.add_document(content="Your text here", topic="general")
Query: result = await rag.query("What is AgentFlow?"); print(result.answer)

Frequently Asked Questions about rag

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How does retrieval-augmented generation reduce LLM hallucinations?▼

Retrieval-augmented generation grounds LLM responses by retrieving relevant documents from a knowledge base and passing that context to the LLM, ensuring answers are based on external factual sources rather than model parameters.

How do I build a document QA system for an internal knowledge base?▼

You can build a document QA system by ingesting text into the knowledge base, performing semantic search to fetch relevant passages, and generating context-aware answers with an LLM using the retrieved document context.

Can I use a provider-agnostic interface to connect my LLM and vector search components?▼

Yes, the Skill integrates memory stores, vector search, and LLM providers via a provider-agnostic interface, allowing you to connect different components without being locked into a specific vendor.

What is the best way to track sources for AI-generated answers in a knowledge base?▼

The best way to track sources for AI-generated answers is using built-in source tracking and topic management features, which provide auditable results by linking responses back to their originating retrieved documents.

Does retrieval-augmented generation work for domain-specific information retrieval?▼

Yes, retrieval-augmented generation is designed for domain-specific information retrieval, fetching relevant passages from ingested documents to answer queries about product documentation, policies, or customer support topics.

What are the limitations of using semantic search for knowledge-base QA?▼

Semantic search for knowledge-base QA depends entirely on ingested documents, meaning it cannot answer queries outside its knowledge base scope and requires documents to be properly added and topic-tagged beforehand.