Code Knowledge Graph in Action: Impact Analysis Demo

Want to know exactly what breaks before you delete a function? This is the structural query RAG can't answer. A code knowledge graph can -- and here's the 70-second live demo of it answering one. In this video, an engineer in the OutcomeOps chat asks "What would break if I remove validateCardToken?" The chat router classifies the query as wiring plus context, fires three MCP tools, and returns a full impact map: 3 direct callers in payments / refunds / admin, 8 transitive callers across handlers and workers, and the ADR that mandates the function as the PCI redactor entry point. The engineer has the full picture before touching a file. This is Video 08 in Series 2: Context Engineering Patterns. Pairs with Video 07 (RAG vs Code Knowledge Graph: Why You Need Both). Chapters 0:00 An engineer wants to remove a function 0:03 Opening OutcomeOps chat 0:09 The classifier reads the query 0:16 Three MCP tools fire: graph + RAG 0:29 Eleven callers across three services 0:34 Graph fanout: hub, direct, transitive 0:45 ADR-031: the PCI redactor entry point 0:57 Wiring plus context 1:05 Real impact analysis. Not a guess. What you will learn - How the chat router classifies queries: application - RAG, specific symbol - Graph, wiring + context = both - What an MCP graph_callers tool call looks like in production - How a code knowledge graph builds a transitive caller fanout (hub, depth=1, depth=2) - How the RAG layer adds the "why" (ADR-031, PCI compliance, SAQ-D re-attestation) - Why "wiring plus context" is the question engineers actually have Why this matters at enterprise scale Removing a function in a Fortune 500 codebase is a multi-million-dollar decision. Pure RAG returns fuzzy semantic matches. Pure greps miss inheritance and dynamic dispatch. A code knowledge graph gives ground truth from the AST. Combined with the design pattern context from RAG, the engineer can make the change with confidence, or know to refactor first. Read more - RAG vs Code Knowledge Graph: https://www.youtube.com/watch?v=Y7J-kki-OlQ - Context Engineering deep dive: https://www.outcomeops.ai/context-engineering - Self-Documenting Architecture: When Code Becomes Queryable: https://www.outcomeops.ai/blogs/outcomeops-self-documenting-architecture-when-code-becomes-queryable - OutcomeOps, the platform: https://www.outcomeops.ai - Brian Carpio (founder): https://briancarpio.com #CodeKnowledgeGraph #GKG #ImpactAnalysis #RAG #ASTAnalysis #ContextEngineering #OutcomeOps