Authorization Gap

Capability vs Authorization

Capability vs. authorization is the distinction between what an AI agent is technically able to do (its capabilities) and what it is actually permitted to do (its authorizations)—two concepts that MCP and agent frameworks do not separate by default.

Why this matters for AI agents

Capability without authorization is dangerous. An agent that can see and call any tool in its catalog may do so regardless of whether the call is appropriate, sanctioned, or safe.

How AgntID relates

AgntID enforces the separation of capability and authorization—an agent may have technical access to a tool, but AgntID's policy evaluation decides whether the agent is authorized to invoke it for the current task.

The static IAM gap

IAM tries to limit capabilities by assigning roles. But for agents, capabilities are discovered at runtime via MCP. IAM roles cannot anticipate which tools an agent will encounter.

Related phrases

capability vs authorizationtool capability vs permissionagent capability gap

Related terms

Frequently asked questions

What is the difference between capability and authorization for AI agents?

Capability is what an agent is technically able to do. Authorization is what it is actually permitted to do. MCP and agent frameworks provide capability; AgntID provides authorization.

Why is capability without authorization dangerous?

An agent that can see and call any tool in its catalog may do so regardless of whether the call is appropriate, sanctioned, or safe.

How does AgntID separate capability from authorization?

AgntID enforces the separation. An agent may have technical access to a tool, but AgntID's policy evaluation decides whether the agent is authorized to invoke it for the current task.

Runtime Enforcement

Secure every agent tool call at execution time.

AgntID gives infrastructure teams scoped, ephemeral access control for AI agents without replacing IAM, MCP servers, tools, or agent frameworks.