Text this: Benchmark Evaluation of a Tool-Augmented Large Language Model Agent Using Traditional Asian Medicine Metadata.