Skip to content

wordlift/wordlift-gemini-cli-extension

Repository files navigation

WordLift Knowledge Graph Builder & Gemini CLI Extension

This extension empowers Gemini with the WordLift Knowledge Graph Builder skill, allowing you to build and maintain Knowledge Graphs directly from your CLI.

🧠 Agent Skills

This extension includes the wordlift-kg-builder skill. When enabled, Gemini gains specialized expertise to:

  • Import Sitemaps: Ingest thousands of pages from sitemap.xml.
  • Build Entities: Create valid schema.org Product, Organization, and WebPage entities.
  • Generate GS1 IDs: Create deterministic, GS1 Digital Link compliant identifiers.
  • Validate Data: Ensure data quality with built-in SHACL validation.
  • Sync to WordLift: Orchestrate batch updates and incremental patches.

To use it, simply ask Gemini:

"Help me import pages from my sitemap" "Build a Product entity with GS1 Digital Link" "Validate this JSON-LD against your SHACL shapes"

Installation

Install the extension directly from GitHub:

gemini extensions install https://github.com/wordlift/wordlift-gemini-cli-extension.git

Configuration

Create a .env file in the extension directory with your WordLift credentials:

WORDLIFT_API_KEY=your_api_key_here
WORDLIFT_BASE_URI=https://data.wordlift.io/wl123
WORDLIFT_API_ENDPOINT=https://api.wordlift.io

Core Features (via MCP)

Even without activating the full Skill, the extension provides basic tools via the Model Context Protocol (MCP):

  • import_from_sitemap: Import URLs from sitemap.xml.
  • build_product: Create schema.org Product entities.
  • validate_entity: Validate entities against SHACL shapes.
  • sync_kg: Perform batch create/update operations.

Architecture

  • skills/: Contains the wordlift-kg-builder Agent Skill.
  • wl_extension/: Core Python logic (entity building, validation, sync).
  • server.py: MCP server implementation.

Support

For WordLift API documentation, visit: docs.wordlift.io

About

Official Gemini CLI extension for WordLift. Manage your Knowledge Graph with natural language using Model Context Protocol (MCP)

Topics

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors