Researchers introduced EvoOntology, a self-evolving ontology layer for data agents. It bridges the agent-data gap for heterogeneous data like tables and databases. The system uses an MCP server with schema, content and tool layers. A builder agent constructs the ontology autonomously. A self-evolution loop refines it via attribution-guided edits. Edits are accepted only after paired evaluation. Tests on four LLM backbones show consistent gains over baselines. Code is available on GitHub.