What problem does it solve? Exploring the Divalto ERP standard codebase is slow when you don't know which programs, functions, or tables are relevant to a change request. This Skill queries the diva-mcp Neo4j graph (X.12 snapshot) to produce a prioritized candidate list before verification in the live X.13 sources. ## Core Features & Use Cases - Cypher Query Generation: Reads a parsed request.json and selects from 7 parameterized Cypher templates (by keyword, by module, table access, static/dynamic/XMT callers, similar entities) with enforced LIMIT bounds. - Result Consolidation: Deduplicates and scores raw Neo4j rows into a candidates_x12.json file with programs, functions, tables, entities, and call relations, all marked with mandatory X.12 disclaimers. - Graceful Degradation: When the diva-mcp server is unavailable, returns an empty candidate report with neo4j_status unavailable so the X.13 source-search phase can take over. - Use Case: Given a feature request about a Retail payment family (RTFTAB), generate the Cypher queries, execute them via the MCP read tool, and consolidate results into a scored candidate list for verification. ## Quick Start Run the query_neo4j.py script in generate mode with your request.json, execute the returned Cypher via the diva-mcp read tool, then rerun the script in consolidate mode with the raw results.
