bicameral-context-sentry

Knowledge graph reconciliation searches existing context to relate incoming decisions, returning a structured report with matches, conflicts, and context signals. It ensures new decisions are grounded in established context before flowing into the pipeline.

Human-in-the-loop probing generates targeted questions when graph reconciliation detects ambiguities that cannot be resolved automatically. It collects user answers through Phase 1 and Phase 2 protocols before handing control back to the calling flow.

To ensure alignment, the context sentry evaluates relationships in the knowledge graph and surfaces targeted questions when new concepts need naming relative to existing context. This grounds decisions in established context before ingestion.

Yes, graph search is used to relate incoming material to existing context and returns a structured reconciliation report highlighting matches, conflicts, and context signals. This identifies misalignments before new decisions enter the pipeline.

Use automatic graph reconciliation first to search existing context and evaluate relationships. Apply human-in-the-loop probing only when ambiguities are detected that cannot be resolved automatically, queuing resolutions for user review.

Automatic context reconciliation cannot resolve all ambiguities between new decisions and existing context. When relationships cannot be evaluated automatically, it must surface targeted questions and queue resolutions for human review before proceeding.