Scenario-Agnostic Zero-Trust Defense with Explainable Threshold Policy: A Meta-Learning Approach | IEEE Conference Publication | IEEE Xplore