Trust inversion and its exploitation: A threat model of the Model Context Protocol for agentic AI. (2026). Journal of the Nigerian Society of Physical Sciences, 8(4), 3748. https://doi.org/10.46481/jnsps.2026.3748