This skill provides a dual approach to graph engineering, covering both the structural design of knowledge graphs and the operational logic of task graphs. It guides the creation of ontologies, entity extraction pipelines, and GraphRAG memory systems while defining patterns for agent orchestration such as parallel fan-out and verifier separation.
Designed for knowledge engineers and automation builders, the skill transforms raw text into structured entity-relation networks and organizes complex agent interactions into manageable, gated workflows.
Follow the documented setup, then try a first task.
Installs: Complete graph-engineering skill · Claude Code
In a terminal, clone the graph-engineering repository.
git clone https://github.com/codejunkie99/graph-engineering.git
In the same terminal and from the directory where the repository was cloned, copy its graph-engineering skill folder into Claude Code's skills directory.
cp -r graph-engineering/graph-engineering ~/.claude/skills/
Uses graph-engineering
Use the graph-engineering skill to build a knowledge graph from my documents at <documents-path>. Start by scoping the domain and proposing the initial schema before extraction.
Uses graph-engineering
Use the graph-engineering skill to teach me graph engineering using <project-or-dataset> as the running example. Begin with the first pipeline stage, include a small diagram, and finish this exchange with one exercise before continuing.
Claude