How does this skill handle citations?

It provides specific patterns to parse 'grounding_metadata' from model responses, extracting source URIs and text segments to show users exactly where information came from.

What parameters can I tune for retrieval?

You can configure 'top_k' for the number of chunks to retrieve and 'vector_distance_threshold' to control the strictness of semantic similarity.

Does this require a Google Cloud project?

Yes, this skill is specifically designed for developers working within the Google Cloud Platform and utilizing Vertex AI services.

What is the Vertex AI RAG Engine?

It is a managed service on Google Cloud that handles the ingestion, embedding, indexing, and retrieval steps required to ground LLM responses in private data.

Can I use local files with this RAG pattern?

Yes, the skill includes patterns for uploading local files to a Google Cloud Storage bucket and then importing them directly into a Vertex AI RAG corpus.

Vertex AI RAG Patterns

Name: Vertex AI RAG Patterns
Author: Agentient

byAgentient

数据科学与机器学习

Implements Vertex AI RAG Engine patterns for grounding AI agents with private data sources and verifiable citations.

关于

This skill provides a comprehensive suite of implementation patterns for the Vertex AI RAG Engine, enabling developers to ground Claude's responses in authoritative, private datasets. It covers the entire five-stage RAG pipeline—from corpus management and document ingestion via Google Cloud Storage to configuring advanced retrieval tools and extracting verifiable citations. By integrating these patterns, developers can significantly reduce hallucinations and build production-ready agents capable of answering complex, domain-specific queries with high precision and transparency within the Google Cloud ecosystem.

主要功能

Robust citation extraction and grounding metadata parsing
Semantic chunking and embedding model configuration patterns
Advanced retrieval tool configuration with tunable similarity thresholds
Automated Vertex AI RAG corpus creation and lifecycle management
Multi-source document ingestion from local files and Google Cloud Storage
0 GitHub stars

使用场景

Developing research tools that provide verbatim citations from academic papers and PDFs
Building a corporate knowledge base assistant for internal policy and HR queries
Creating customer support bots grounded in technical product manuals and documentation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add agentient/vibekit rag-patterns

For use in Claude.ai and ChatGPT

Download Skill

GitHub

关于