AI systems are no longer isolated tools; they increasingly operate as connected components that request access, exchange context, and act across services and environments. The Model Context Protocol (MCP) is emerging as a key mechanism for how AI systems discover, negotiate, and use context, with significant implications for identity, trust, and governance.

This hackathon-style workshop gives Identiverse 2026 attendees a practical grounding in the current state of MCP and what it means for identity professionals. Participants will start with a clear, implementation-oriented overview of MCP as it exists today and then focusing on how context is actually represented, shared, and consumed by AI-enabled systems.

Attendees will then take part in guided, hands-on lab activities that explore:

• How AI systems use MCP to request and apply contextual information
• Where identity signals appear, implicitly and explicitly, in MCP-based interactions
• How trust boundaries shift when context is dynamically assembled and reused
• What new risks and design considerations emerge for enterprise and consumer environments

This workshop offers pragmatic education, direct interaction with real systems, and an emphasis on understanding behaviour rather than theory. Participants will leave with a clearer mental model of MCP, practical experience engaging with it, and a stronger sense of how AI identity is shaped by context-driven architectures.