NLWeb Schema.org Grounding
Structures and audits website content as Schema.org JSON-LD to enable high-fidelity data retrieval for AI agents.
This skill facilitates the preparation and validation of structured data for NLWeb ingestion, ensuring that web content is correctly mapped to Schema.org JSON-LD formats. It allows developers to audit existing sites, author new structured data, and manage site-specific configurations in sites.xml. By grounding data in formal schemas, it enables AI agents to receive fully structured objects—such as products with prices or recipes with ingredients—rather than generic text chunks. This significantly improves the accuracy and actionability of agentic commerce applications and complex tool routing systems.
主な機能
01Optimizes the schema_object flow for high-fidelity agent responses
02Manages site-specific registration and tool inheritance in sites.xml
03Audits existing website metadata for NLWeb ingestion compatibility
04Structures content into Schema.org JSON-LD for types like Product, Recipe, and Article
05Ensures data follows ISO standards for dates and absolute URL conventions
0632 GitHub stars
ユースケース
01Auditing structured data on content sites to improve RAG performance
02Preparing e-commerce product data for integration with AI shopping assistants
03Configuring domain-specific tool routing for niche site types like RealEstate or Courses