searching

Retrieve cited answers from documents and knowledge graphs with hybrid search.

36|18|Updated Nov 23, 2025

Search Tags:#summarization #knowledge graph #search #citation #hybrid retrieval #web fallback #document QA

Ready to run

npx skills add https://github.com/aws-samples/sample-aws-idp-pipeline --skill searching-aws-samples

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: searching
Source: https://github.com/aws-samples/sample-aws-idp-pipeline/tree/main/packages/agents/idp-agent/.skills/search
Command: npx skills add https://github.com/aws-samples/sample-aws-idp-pipeline --skill searching-aws-samples

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps you quickly locate specific information inside uploaded documents and the knowledge graph when you do not know where the answer is stored.

Core Features & Use Cases

Hybrid document search to retrieve relevant passages with sources for factual answers.
Knowledge-graph traversal and keyword lookup to expand from what you found into related entities and concepts.
Optional web lookup fallback when document search cannot cover the request well.

Use Case Example: You ask, “What policies apply to expense reimbursements for contractors, and what exceptions are mentioned?” and the skill searches your documents, follows related graph connections, and returns a sourced explanation.

Quick Start

Ask: “Search my uploaded materials for the section that answers my question about expense reimbursements for contractors and summarize the key exceptions with sources.”

Frequently Asked Questions about searching

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

FAQPage Schema

How do I find specific answers across my uploaded documents with citations?▼

Document search uses hybrid retrieval to locate relevant passages and returns factual answers with source citations. It scans your uploaded materials to extract precise information and provides references for verification.

Can I use knowledge graph traversal to expand document search results?▼

Knowledge graph traversal expands document search by connecting related entities and concepts. You can perform keyword lookups alongside hybrid retrieval to discover linked information across graph-connected nodes.

Does web search fallback work when document search is insufficient?▼

Web search fallback activates optionally when document search cannot adequately cover your request. It fetches external sources and returns results with URLs to supplement missing information from your uploaded documents.

What's the best way to summarize and compare information across document collections?▼

Summarization and comparison across document collections is done through hybrid retrieval combined with graph traversal. The skill extracts relevant passages, identifies key exceptions, and returns sourced explanations for comparative analysis.

Why does hybrid retrieval return results without finding the specific policy section I need?▼

Hybrid retrieval may miss specific sections if document indexing lacks sufficient keyword density or graph connections. Use more targeted query terms or enable web fallback to supplement insufficient document coverage.