??A flaky test is a test that its outcome varies across executions without any intentional change to the source code or test logic. Such non-determinism arises from a range of root causes, including order dependencies between tests, non-idempotent side effects on shared state, reliance on timing or external resources, and genuine non-determinism in the system under test. ??Regardless of origin, flaky tests impose significant costs on software teams: they erode confidence in test suites, inflate debugging effort, and complicate continuous integration pipelines by introducing false signals that obscure real regressions. Despite the research attention, existing automated approaches remain largely static and category-specific, leaving a clear gap for agentic workflows that can dynamically gather execution evidence, reason over it, and iteratively refine repairs across diverse flakiness types.