Query the RawgrowthOS knowledge graph for grounded answers via HTTP API.
npx skills add https://github.com/scanbott/claude-skills --skill rag-query-scanbott
Please help me install this Agent Skill. Skill: rag-query Source: https://github.com/scanbott/claude-skills/tree/main/rag-query Command: npx skills add https://github.com/scanbott/claude-skills --skill rag-query-scanbott
Answers grounded in indexed documents can be hard to retrieve from the RawgrowthOS knowledge graph. This Skill provides a direct way to query the graph and obtain answers tied to relevant documents.
Just provide a question to the RAG service endpoint to get a grounded answer.
High-intent search queries and answers about installing and using this skill.
To retrieve grounded answers from a knowledge graph, send a JSON payload with your question and querying mode to the HTTP API endpoint. The RAG service returns a structured answer tied to indexed documents.
Hybrid mode fits general querying, local mode targets specific entity lookups, global mode uncovers cross-domain patterns, and naive mode performs simple vector similarity retrieval from indexed documents.
Yes, you can query the knowledge graph using a simple HTTP curl-based workflow. It accepts a JSON payload and returns clear response formats, enabling lightweight integration with your existing tools.
You need global mode when you want to uncover cross-domain patterns within the knowledge graph. It queries indexed documents to return grounded answers spanning multiple domains, unlike specific local lookups.
No, the RAG query API operates without external dependencies. It uses a lightweight curl-based workflow to query indexed documents and retrieve structured answers directly from the knowledge graph.
Alex1980AlexIndex PDFs into vector and graph knowledge stores for search and extraction.
itsmarsssExport an Obsidian vault's wikilink graph to JSON, GraphML, Cypher, and HTML.
kyssta-exeIngest financial PDFs, Excel, and CSV reports into a searchable knowledge graph.
Kingly-AgencyBuild RAG pipelines that ground LLM responses with retrieved external knowledge.
