MCP Server Discovery and Registry Design for Large Organizations
Enterprises need separate discovery and authorization layers to govern MCP server access safely.
Declan Mwangi-Carroll
Contributing Editor, Attack Surfaces & Adversarial AI
Declan Mwangi-Carroll
Declan is a former red-team consultant whose work on adversarial prompt manipulation was cited in two industry security advisories before he turned to full-time writing in 2022. He leads Agentic Review's coverage of offensive and defensive research at the boundary of language models and autonomous pipelines.
11 stories
MCP Server Lifecycle Governance From Deployment to Decommission
Securing MCP servers requires governance across their entire lifecycle, not just at deployment.
Declan Mwangi-Carroll
ABAC vs RBAC for AI Agent Permission Models
ABAC's context-aware policies solve RBAC's inability to handle dynamic agent permissions.
Declan Mwangi-Carroll
A2A Protocol and MCP Interoperability in Enterprise Agent Architectures
MCP connects agents to tools; A2A connects agents to each other in enterprise workflows.
Declan Mwangi-Carroll
MCP Gateway Vendors Compared on API Governance and Policy Enforcement
Enterprise teams deploy AI agents 150,000 times faster than governance infrastructure can scale.
Declan Mwangi-Carroll
OAuth 2.0 and SSO Support Across MCP Control Planes
Enterprise identity systems weren't built for autonomous agents making API calls.
Declan Mwangi-Carroll
RBAC Design for AI Agents in Enterprise Systems
Traditional access controls weren't built for machines that operate autonomously at scale.
Declan Mwangi-Carroll
Multi-Agent Control Plane Design Patterns
Governance frameworks for controlling sprawling agent deployments across enterprise systems.
Declan Mwangi-Carroll
AI Data Protection Obligations Under GDPR and CCPA
Regulators are targeting AI systems that handle personal data under stricter penalties.
Declan Mwangi-Carroll
Securing AI Coding Assistants in Developer Environments
Organizations must govern AI coding assistants before secrets and vulnerabilities proliferate.
Declan Mwangi-Carroll
PII Detection and Blocking in AI Agent Outputs
Enterprises leak PII through AI at scale; hybrid detection and policy enforcement can stop it.
Declan Mwangi-Carroll