11mirror-graph-reasoning

Knowledge graph routing directs relationship questions to the graph engine while retrieving content via brain search. The decision protocol mandates attempting brain search first, then using graph queries for entity connections and provenance.

Use graph_query when investigating entity relationships, provenance, or connections across people, vendors, and cost centers. Brain search handles direct content retrieval, while graph traversal maps structural links between documents and entities.

The graph reasoning decision protocol requires attempting brain search first for content. If relationship mapping is needed, route to graph_query, using tools like entity_search and traverse_graph, while never skipping the initial brain search step.

Traverse the knowledge graph using graph_query and traverse_graph tools to find provenance and connections. Start with a brain search to retrieve base content, then apply graph traversal to map relationships across people, vendors, and documents.

Yes, the graph reasoning protocol enforces a strict decision rule: attempt brain search first for content retrieval, then use graph_query for relationship mapping. This ensures source attribution and prevents skipping content discovery.

Graph queries alone miss direct content retrieval context. The protocol enforces brain search first because skipping it loses source attribution and document-level content before mapping entity relationships and connections via graph traversal.