biocypher

Build a knowledge graph from data files using biotope, the CLI for the BioCypher ecosystem (Croissant-described data → BioCypher graph, with git-like metadata version control). Use this whenever the task is to turn tables, CSVs, JSON, parquet, or mixed/unstructured sources into a queryable knowledge graph or graph database (Neo4j, DuckDB) — including ingesting a dataset into a graph, mapping fields to entities and relations, building a BioCypher project, connecting biomedical/biological data across sources, or grounding data against ontologies. Trigger on mentions of biotope, BioCypher, Croissant, knowledge graph / KG construction, graph ingestion, "turn this data into a graph", entity/relation mapping, or ontology-grounded data integration — even when the user names the goal (a graph) rather than the tool.

Connect biomedical LLM chat to knowledge graphs, APIs, and document RAG with BioChatter. Use for natural-language Cypher over Neo4j, BioCypherPromptEngine query generation, DatabaseAgent, or general biomedical conversation with structured backends. Trigger on BioChatter, NL graph queries, Cypher generation, KG chat, schema_info.yaml querying, or conversational AI over biomedical data.

Build knowledge graphs with BioCypher — adapters, schema_config.yaml, biocypher_config.yaml, and multi-backend export (Neo4j, CSV, PostgreSQL, RDF). Use for standalone KG ETL pipelines, writing schema_info.yaml, Neo4j bulk import, or tuning dbms/output settings. Trigger on BioCypher, schema_config, biocypher_config, write_nodes, write_edges, knowledge graph construction, Biolink schema, or graph database export.

Remove signs of AI-generated writing from text and revise drafts so they sound natural, specific, and human-written. Use when editing, reviewing, or rewriting documentation, guides, landing pages, release notes, or other prose to reduce AI tells such as inflated symbolism, promotional language, superficial -ing analysis, vague attributions, em dash overuse, rule-of-three cadence, AI-coded vocabulary, passive voice, negative parallelisms, filler phrases, generic conclusions, and chatbot-like framing.

Project-wide specification and test-driven workflow for every development change in this documentation repository. Use before any feature, fix, refactor, content or UI modification, configuration change, or other implementation work, then resume after implementation to verify the agreed acceptance specs. Align on the user's what and why, formalize minimal MMECE outcomes with planned tests, and validate visual documentation at 16:9, 3:4, and mobile sizes.