For a specified quantum causal graph with hidden systems and restricted interventions, characterize which causal effects are identifiable and derive complete inference procedures and sample-complexity bounds, or prove observational indistinguishability.