RBAC Design for MCP Tool Access in Regulated Enterprises
Dynamic tool discovery breaks static permission models designed for fixed resource inventories.
Colin Reyes
LetterMCP
Agent Security
MCP Tool Call Authorization Models Compared
Five authorization models solve the problem of AI agents calling tools without human configuration.
Unauthorized MCP Tool Calls and Audit Trail Requirements
Enterprises lack visibility into AI tool calls because MCP was built for convenience, not security.
Reuben Weldon
More From LetterMCP
MCP gateway and catalog platforms compared
Security and compliance hang on identity, credentials, and policy—not just features.
Mira Okoro
Claude Constitutional AI Limits Under Adversarial Prompting
Adversarial fine-tuning can slip past Claude's Constitutional Classifiers entirely.
Colin Reyes
What MCP Tool Annotations Tell You About Agent Behavior
Annotations signal risk but don't guarantee tool behavior stays honest.
Reuben Weldon
Role of Platform Engineering in AI Agent Enablement
Platform engineers must rebuild identity, access.
Mira Okoro
Enterprise AI Agent Deployment Patterns
Governance and audit trails separate production agents from prototypes.
Mira Okoro
MCP OAuth and Credential Handling Fundamentals
Most MCP servers still use static keys instead of OAuth, creating widespread security risk.
Colin Reyes
MCP Host vs Client Distinction in Agent Architecture
Mixing up host and client roles in MCP creates security gaps in production.
Ayana Brennan
Model Context Protocol and OpenAI Agents SDK Compatibility
Native compatibility solves tool integration, but authentication and transport gaps remain.
Mira Okoro
Autonomous Agent Roles in MCP Orchestration Workflows
Agents juggling multiple roles create governance gaps MCP doesn't solve.
Colin Reyes