Abstract
Mobile operating systems and applications running on mobile terminals may rely on local autonomous AI-driven engines. These AI engines often require access to edge AI capabilities such as external context feeds or computing resources to execute high-complexity queries from a user/application. Existing architectures cannot support provisioning such dynamic AI capabilities in cellular networks. Techniques for provisioning and authenticating AI agent skills and Model Context Protocol (MCP) endpoints are described herein. The techniques extend the GSMA TS.43 Entitlement Configuration Architecture into an AI-native Capability Entitlement framework, which enables mobile terminals to dynamically discover, authenticate and provision MCP servers and network skill manifests in a secure manner. The techniques are agnostic to specific 3GPP releases and therefore, can be deployed across both legacy and future 3GPP networks without requiring expensive and time-consuming network upgrades.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Barry, Aguibou and Cormier, Jean-Philippe, "Dynamic Provisioning and Authentication of AI Agent Skills and MCP Endpoints via Mobile Device Entitlement Configuration Architecture", Technical Disclosure Commons, (September 16, 2026)
https://www.tdcommons.org/dpubs_series/11761