Discussion
Ontology-informed context is key to better outcomes with AI agents such as Claude, ChatGPT, Grok Bot, Muse, OpenCode, and Pi. That much is clear. The shape and form it should take are less settled.
Here's one answer: LOAM in knowledge graph mode.
#VirtuosoRDBMS #KnowledgeGraph #Ontology #AI #Agentic #DataSpaces #SemanticLayer #ContextLayer
Ask OPAL, Claude, ChatGPT, Grok Bot, Muse or any other agent harness to install agent-rdf-memory and its associated SKILL.md skills into your Virtuoso instance.
Try Virtuoso: https://shop.openlinksw.com/license_generator/virtuoso/
agent-rdf-memory: https://github.com/OpenLinkSoftware/ai-agent-skills/tree/main/agent-rdf-memory
Skills: https://github.com/OpenLinkSoftware/ai-agent-skills
LOAM (Linked Open Agent Memory) keeps an agent's rules in an ontology-informed knowledge graph. The agent checks which rules apply, then reaches for skills first, like Data Twingler for SQL, SPARQL and SPASQL queries. Any model, any harness.
Link: https://linkeddata.uriburner.com/weblog/?post=beyond-llm-wiki-and-query-claude_sonnet_5_5.html