This is a typed knowledge graph that lets Claude store and query structured information as entities and relations. You'd reach for it when you need to remember people, projects, tasks, or documents and how they connect, especially when multiple skills need to share state. Everything validates against a schema before committing, so you can enforce constraints like "tasks must have owners" or "no circular dependencies." The planning-as-graph-transformation pattern is clever: model multi-step work as a sequence of CREATE and RELATE operations that roll back if constraints break. Storage is append-only JSONL by default, which keeps things simple until you need SQLite. The cross-skill integration examples are the real payoff here.
$ npx -y skills add sundial-org/awesome-openclaw-skills --skill ontology --agent claude-codeInstalls into .claude/skills of the current project.
A typed vocabulary + constraint system for representing knowledge as a verifiable graph.
Everything is an entity with a type, properties, and relations to other entities. Every mutation is validated against type constraints before committing.
Entity: { id, type, properties, relations, created, updated }
Relation: { from_id, relation_type, to_id, properties }
| Trigger | Action |
|---|---|
| "Remember that..." | Create/update entity |
| "What do I know about X?" | Query graph |
| "Link X to Y" | Create relation |
| "Show all tasks for project Z" | Graph traversal |
| "What depends on X?" | Dependency query |
| Planning multi-step work | Model as graph transformations |
| Skill needs shared state | Read/write ontology objects |
# Agents & People
Person: { name, email?, phone?, notes? }
Organization: { name, type?, members[] }
# Work
Project: { name, status, goals[], owner? }
Task: { title, status, due?, priority?, assignee?, blockers[] }
Goal: { description, target_date?, metrics[] }
Sponsors