Exécutez n'importe quel Skill dans Manus
en un clic

Exécutez n'importe quel Skill dans Manus en un clic

retrieval-strategies

Étoiles6

Forks0

Mis à jour10 mai 2026 à 11:10

Route RAG retrieval quality work across hybrid search, reranking, query transformation, HyDE, Self-RAG, RAPTOR, CRAG, and Graph RAG.

Installation

Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.

Exécuter dans Manus

Source

Goodnight77

Goodnight77/rag-skills

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Téléchargement

Exécuter dans Manus

Métiers associésSOC

Basé sur la classification professionnelle SOC

Scientifiques des donnéesProfessions informatiques et mathématiques·SOC 15-2051

Explorateur de fichiers

13 fichiers

SKILL.md

readonly

name	retrieval-strategies
title	Retrieval Strategies
description	Route RAG retrieval quality work across hybrid search, reranking, query transformation, HyDE, Self-RAG, RAPTOR, CRAG, and Graph RAG.
category	retrieval-strategies
tags	["retrieval","ranking","hybrid-search","rag"]
allowed-tools	["Read","Grep","Glob"]

Retrieval Strategies

Overview

Use this parent skill when the main RAG problem is search quality, ranking, recall, context selection, or evidence traceability. Route to the child skill that best matches the retrieval failure.

Problem Statement

Embedding search alone often misses exact terms, ranks weak evidence too high, or retrieves incomplete context. RAG systems need retrieval strategies that adapt to query type and evidence requirements.

Key Concepts

Hybrid search: Combine sparse and dense signals.
Reranking: Reorder broad candidate sets with stronger relevance models.
Query transformation: Rewrite, expand, or decompose user questions.
Corrective retrieval: Validate and repair weak evidence before generation.

Implementation Guide

Step 1: Diagnose the Failure

Identify whether the issue is recall, ranking, query mismatch, missing context, or poor citations.

Step 2: Select the Retrieval Pattern

Choose hybrid search for exact-term misses, reranking for noisy top-k results, query transformation for vague questions, and corrective patterns for unreliable evidence.

Step 3: Evaluate With Grounded Metrics

Measure recall at k, MRR or nDCG, citation accuracy, answer faithfulness, and latency impact.

When to Use This Skill

Improving retrieval quality in an existing RAG pipeline
Selecting a search pattern for a new RAG system
Debugging weak citations or irrelevant context

When NOT to Use This Skill

The source content has not been parsed or chunked yet
The main issue is vector database deployment
The task is pure generation without retrieval

Related Skills

Metrics & Success Criteria

Improved recall and ranking quality on evaluation queries
Better answer faithfulness and citation accuracy
Latency remains acceptable for the product workflow

Plus depuis ce dépôt

même dépôt

rag-skills

Goodnight77/rag-skills

Use this skill when building, debugging, or improving Retrieval-Augmented Generation systems, including chunking, vector database selection, hybrid search, reranking, multimodal RAG, code documentation RAG, retrieval latency, and production RAG architecture.

2026-07-106

hierarchical-chunking

Goodnight77/rag-skills

Chunk nested documents into parent-child levels so retrieval can move from broad sections to fine-grained passages.

2026-05-106

semantic-chunking

Goodnight77/rag-skills

Use semantic boundaries and embedding similarity to chunk text for higher-relevance retrieval.

2026-05-106

chunking

Goodnight77/rag-skills

Route RAG chunking decisions across semantic, hierarchical, sliding-window, contextual-header, and framework-selection strategies.

2026-05-106

sliding-window-chunking

Goodnight77/rag-skills

Use overlapping windows to preserve context across chunk boundaries while controlling retrieval size.

2026-05-106

optimize-retrieval-latency

Goodnight77/rag-skills

Reduce retrieval latency with caching, batching, and index-level optimization.

2026-05-106

name	retrieval-strategies
title	Retrieval Strategies
description	Route RAG retrieval quality work across hybrid search, reranking, query transformation, HyDE, Self-RAG, RAPTOR, CRAG, and Graph RAG.
category	retrieval-strategies
tags	["retrieval","ranking","hybrid-search","rag"]
allowed-tools	["Read","Grep","Glob"]

Retrieval Strategies

Overview

Use this parent skill when the main RAG problem is search quality, ranking, recall, context selection, or evidence traceability. Route to the child skill that best matches the retrieval failure.

Problem Statement

Key Concepts

Hybrid search: Combine sparse and dense signals.
Reranking: Reorder broad candidate sets with stronger relevance models.
Query transformation: Rewrite, expand, or decompose user questions.
Corrective retrieval: Validate and repair weak evidence before generation.

Implementation Guide

Step 1: Diagnose the Failure

Identify whether the issue is recall, ranking, query mismatch, missing context, or poor citations.

Step 2: Select the Retrieval Pattern

Choose hybrid search for exact-term misses, reranking for noisy top-k results, query transformation for vague questions, and corrective patterns for unreliable evidence.

Step 3: Evaluate With Grounded Metrics

Measure recall at k, MRR or nDCG, citation accuracy, answer faithfulness, and latency impact.

When to Use This Skill

Improving retrieval quality in an existing RAG pipeline
Selecting a search pattern for a new RAG system
Debugging weak citations or irrelevant context

When NOT to Use This Skill

The source content has not been parsed or chunked yet
The main issue is vector database deployment
The task is pure generation without retrieval

Related Skills

Metrics & Success Criteria

Improved recall and ranking quality on evaluation queries
Better answer faithfulness and citation accuracy
Latency remains acceptable for the product workflow