The engine runs a structured model-based evaluation pass to determine whether live research findings corroborate or refute recalled claims, updating the claim's status while attaching retrieved source citations.
flowchart LR
A[Research Findings] --> C{LLM Evaluator}
B[Parametric Claims] --> C
C -->|Corroborates| D[Status: Confirmed]
C -->|Refutes| E[Status: Contradicted]
C -->|Unmatched Finding| F[Status: Memory Gap]
D --- G[(Attach Citations)]
E --- GReferenced by