skills.rest
yonatangross
yonatangrossCOMMUNITY
Explore all skills by this author →
𝕏@yonyonizGitHub

rag-retrieval

Implement RAG patterns with hybrid search, reranking, and PGVector integration.

217|20|Updated Dec 31, 2025
Search Tags:#embeddings#llm#rag#retrieval#grounding#hybrid search#pgvector
One-click install
npx skills add https://github.com/yonatangross/orchestkit --skill rag-retrieval
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: rag-retrieval
Source: https://github.com/yonatangross/orchestkit/tree/main/plugins/ork/skills/rag-retrieval
Command: npx skills add https://github.com/yonatangross/orchestkit --skill rag-retrieval

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill addresses the challenge of LLM hallucinations and lack of factual grounding by implementing advanced Retrieval-Augmented Generation (RAG) techniques. It ensures LLM responses are accurate, verifiable, and based on provided context.

Core Features & Use Cases

  • Core RAG: Basic retrieval, context assembly, and citation generation.
  • Hybrid Search: Combines semantic (vector) and keyword (BM25) search for comprehensive coverage.
  • Advanced RAG Patterns: Implements HyDE, Agentic RAG, Multimodal RAG, Query Decomposition, and Reranking.
  • Vector Database Integration: Supports PGVector for efficient hybrid search.
  • Use Case: Building a customer support chatbot that answers questions based on product documentation, ensuring all answers are cited and factually accurate.

Quick Start

Use the rag-retrieval skill to implement a RAG pipeline that answers questions with inline citations.

Frequently Asked Questions about rag-retrieval

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

FAQPage Schema
How do I implement retrieval-augmented generation to ground LLM responses?▼

Retrieval-augmented generation grounds LLM responses by implementing retrieval pipelines that assemble context and generate citations. This ensures answers are factually accurate, verifiable, and based on provided documentation to reduce hallucinations.

What is hybrid search and how does it improve RAG retrieval quality?▼

Hybrid search improves RAG retrieval quality by combining semantic vector search with BM25 keyword search. This dual approach provides comprehensive coverage, ensuring both precise term matches and broad semantic context are retrieved.

Can I use pgvector for vector database integration in my RAG pipeline?▼

Yes, pgvector is supported for vector database integration to enable efficient hybrid search. It allows you to store and query embeddings directly within your database alongside your relational data for streamlined retrieval.

How do I reduce LLM hallucinations when building a customer support chatbot?▼

To reduce LLM hallucinations in a customer support chatbot, use RAG patterns to answer questions based strictly on product documentation. This contextual retrieval enables citation tracking for verifiable, factually grounded answers.

What advanced RAG patterns are available for query decomposition and reranking?▼

Advanced RAG patterns available include HyDE, agentic RAG, multimodal RAG, query decomposition, and reranking. These techniques optimize retrieval quality by refining search queries and reordering results for maximum contextual relevance.

Does this RAG implementation support multimodal retrieval and agentic workflows?▼

Yes, the implementation supports multimodal RAG and agentic RAG workflows. These advanced patterns allow the system to process multiple data modalities and autonomously decide when to retrieve context during complex reasoning tasks.

Related Skills

Explore More Software Engineering Skills→
Mercurium-AnalyticsMercurium-Analytics

django-pgsearch-patterns

Integrate Django with pg_search and pgvector for hybrid BM25 and vector search.

Official
Advanced
Agents4Academia-AIAgents4Academia-AI
9

ask

Synthesize cited answers from local research papers and notes.

Official
Intermediate
NotysotyNotysoty
9

Hybrid Search Architect

Design hybrid retrieval pipelines combining dense vector search with BM25 keyword matching.

Community
Advanced
voidrotvoidrot

similarity-search-patterns

Implement efficient similarity search using vector databases.

Community
Advanced