ddkg.skill: A Compositional Agent Skill for Translating Biomedical and Bioinformatics Questions into Cypher for the Data Distillery Knowledge Graph

ddkg.skill enables accurate biomedical and bioinformatics queries on the NIH DDKG with 38 files and 239 routing relationships, achieving correct results in 5 of 7 tests.

• Why it matters: Biomedical knowledge graphs are complex and difficult to query correctly without detailed schema knowledge, limiting effective data integration and analysis.
• What they did: The team developed ddkg.skill, an Agent Skill with a controller, structured references, documentation, validated query patterns, and a Python script, tested on an earlier build with 38 files and 227 routing entries.
• The result: In a nine-test evaluation against a live December 2025 DDKG instance, the skill produced correct results in most tests, demonstrating its potential to guide AI in complex biomedical data querying and integration.