What problem does it solve? Neo4j graph results reflect an older X.12 snapshot of the Divalto ERP and may not match the current X.13 source tree. This Skill confirms which candidate programs still exist on disk, flags disappeared ones, and discovers new relevant code, producing a trustworthy evidence file instead of stale graph data. ## Core Features & Use Cases - Candidate Verification: Glob-searches the ERP tree for each Neo4j candidate program and promotes it to CONFIRME X.13 or DISPARU X.13 status. - Context Extraction: Parses DIVA source structure (Procedure/Function blocks) to return the enclosing block plus a 10-30 line snippet for each match. - Bounded Search & Impact Analysis: Applies strict limits (--max-matches, --max-files, --timeout) across ~7000 files, greps technical keywords for new findings, and detects callers via Call/Execute patterns. - Use Case: After querying the diva-mcp graph for candidates related to a Retail feature request, run the verification workflow to confirm which programs exist in X.13, extract their relevant procedures, and list their callers for an impact study. ## Quick Start Ask the AI to verify the Neo4j candidates in candidates_x12.json against the local ERP source tree and produce an evidence_x13.json report with confirmed, disappeared, and newly discovered programs.
