Trust Inversion and Its Exploitation: A Threat Model of the Model Context Protocol for Agentic AI. J. Nig. Soc. Phys. Sci. 2026, 8 (4), 3748. https://doi.org/10.46481/jnsps.2026.3748.