Skip to main content
Arif Mughal

Tag

#MCP

1 article with this tag.

AI Security

Agentic AI Security Architecture: Securing Autonomous Agents in the Enterprise

A research-grounded enterprise reference architecture for autonomous AI agents: agent identity, task-bound authorization, tool and MCP gateways, memory security, runtime containment, observability, and incident response — with an OWASP agentic Top 10 mapping, a bounded-autonomy model, and a 38-control matrix.

71 min read