ontology

This skill provides a typed ontology system for representing structured knowledge as a verifiable graph. It is designed to give AI agents and composable skills a consistent way to store, retrieve, and update shared memory using strongly typed entities and explicit relationships. Each piece of information is modeled as an entity with an ID, type, properties, and relations, ensuring that all stored knowledge follows a predictable and constraint-validated structure. Every change to the graph is validated against predefined type rules before being committed, which helps prevent inconsistent or invalid state.

The system supports core capabilities such as entity creation, querying, updating, and relationship management, along with graph traversal operations like dependency resolution and cross-entity lookups. It includes a schema-based constraint layer defined in a YAML configuration file, allowing developers to enforce required fields, enum values, forbidden properties, and relational rules such as cardinality or acyclicity. Data is stored in a JSONL-based graph format by default, with the option to migrate to SQLite for larger or more complex graphs. A set of scripts (ontology.py) provides command-line operations for creating, querying, linking, retrieving, and validating entities, while a skill contract mechanism defines what data a skill reads or writes and its execution constraints.

This skill is intended for AI systems that require persistent, structured memory and explicit reasoning over interconnected data. Common use cases include task tracking, project planning, dependency management, linking people and organizations, and enabling multi-step reasoning through graph transformations. It is especially useful for agent frameworks where multiple skills must share consistent state or coordinate actions. Developers, AI system architects, and multi-agent application builders benefit from its ability to maintain structured, validated, and queryable knowledge across workflows.