Give an AI agent bounded Cypress Test Replay evidence, compare flaky attempts, and verify a proposed fix without exposing unnecessary access.