ontology

To build a typed knowledge graph for agent memory, use this Skill to model entities, properties, and relationships for verifiable memory. It applies to People, Projects, Tasks, Events, and Documents, enabling CRUD operations and relation linking.

Agents share state across skills through a local memory store at memory/ontology. This Skill uses a stored ontology to enable cross-skill data sharing and lightweight reasoning by linking related typed objects.

Constraint validation for knowledge graph entities works by checking mutations against type constraints defined in a schema file. This Skill enforces an append-only history to preserve provenance and ensure verifiable memory.

Yes, you can plan multi-step workflows as graph transformations. This Skill provides graph operations for linking, querying, and planning actions across skills to execute multi-step workflows.

To run graph validation for entity relations, you need a local memory store initialized under memory/ontology and a schema file. This Skill uses a Python CLI to apply append-only mutations and validate against the defined schema.

Use a typed graph memory over unstructured memory when you need verifiable memory with provenance, consistent modeling of specific entities, and cross-skill data sharing. It enforces type constraints and relation linking for reliable knowledge management.