ontology

You can create and query typed entities in a knowledge graph for agent memory using a Python CLI to manage structured data, validate constraints against a YAML schema, and store records in an append-only JSONL log.

Sharing state and linking related objects across multiple agent skills is managed through a typed ontology graph that enables cross-skill data sharing, enforces constraints, and models multi-step planning as graph transformations.

Schema validation for mutating entities and relations in graph memory works by reading an optional YAML schema definition and enforcing type and constraint rules before applying runtime operations and appending data to the JSONL log.

Yes, you can use a Python CLI to create, list, update, delete, and validate typed entities like Person, Project, and Task, ensuring all operations conform to a defined YAML schema and are tracked in an append-only log.

Yes, YAML is required as a dependency to define and enforce the typed vocabulary and constraint system that validates entities and relations within the ontology graph memory.

An append-only JSONL log stores knowledge graph entities and relations to preserve a verifiable history of mutations, ensuring data consistency and constraint validation across composable agent skills.