KGCS

Knowledge Graph for CyberSecurity

The single source of truth for semantics and contracts that let AI systems reason about cybersecurity deterministically, through an explicit causal chain.

CPE → CVE/CVSS → CWE → CAPEC → ATT&CK → {D3FEND, CAR, SHIELD, ENGAGE}

What's in the spec

Five doors

Pick the door for what you came to do.

Spec — what is KGCS and how is it designed?

For researchers, architects and reviewers who want the causal chain, the architecture and the reasoning behind it before touching any artifact.

Ontology — what can I consume, and under what URI?

For tool builders, pipelines, agents and other ontologies that need a stable, citable identifier for every module. The ontology prefix URIs do not move.

Agents — how does an agent reason over the graph?

For people building agents: the contracts, schemas and traversal templates an agent needs to query the graph and stay within its confidence and validation rules.

Case studies — does it actually work?

For the skeptical, the community and anyone evaluating KGCS: real vulnerabilities walked through end to end, reasoning traces included.

Talks — I want the explained version

For a first contact with KGCS, or if you're sitting in the room while someone presents it.

Source of truth

Version 1.0 is frozen; new semantics ship as versioned successors. Apache 2.0 license, attribution preserved for NIST/MITRE intellectual property. The full spec, changelog, shapes, mappings and contracts live in the official repository — that is where changes, issues and versioned releases happen.

github.com/Ariadna-KGCS/kgcs-spec →