codejunkie99/graph-engineering

What it solves

It addresses the limitation of relying solely on prompt engineering for AI agents by introducing "Graph Engineering." This approach focuses on designing the structural topology of how agents remember information (Knowledge Graphs) and how they execute work (Task Graphs) to improve reliability and scalability.

How it works

The project provides a framework and a set of "skills" that can be integrated into AI agents (like Claude Code). It implements a structured 9-stage pipeline for knowledge graph creation—ranging from scoping and ontology design to fusion and serving to LLMs. For task execution, it applies specific architectural patterns such as the "diamond" (parallel workers with separate verifiers) and the "human gate" to manage complex workflows.

Who it’s for

It is designed for developers and AI engineers who want to move beyond simple prompting and build sophisticated agentic systems using structured knowledge and optimized execution flows.

Highlights

• 9-Stage KG Pipeline: A rigorous process for building knowledge graphs (Scope $\rightarrow$ Representation $\rightarrow$ Ontology $\rightarrow$ Entities $\rightarrow$ Relations $\rightarrow$ Events $\rightarrow$ Quality Gate $\rightarrow$ Fusion $\rightarrow$ Serving).
• Task Graph Patterns: Implementation of research-backed orchestration rules, including parallel fan-out and strategic human-in-the-loop gates.
• Agent-Ready Skills: Packaged skill files and prompt blocks that allow an AI agent to either teach the user graph engineering or actively build a graph from documentation.
• Academic Foundation: Distills a graduate-level Knowledge Graph course from Southeast University and research from Google DeepMind and MIT.