Use cases

These use cases demonstrate how to onboard and secure AI agents and tools using IBM Verify. Each use case is a runnable end-to-end implementation built around a Course agent scenario and shows a distinct integration pattern for securing agent tool invocations on behalf of a signed-in human.

Use case	Description	Link
agent-identity-course-agent-sample	Onboard and secure a conversational AI agent by using direct tool integration. The agent authenticates a human through Authorization code and PKCE, obtains an actor token through Client Credentials, and exchanges them at IBM Verify STS to produce a delegated token. The Course API validates this delegated token before processing any requested operation.	View
agent-identity-course-agent-mcp-sample	Same delegated identity flow, but tools are registered on a separate MCP server and invoked over the Model Context Protocol. The delegated token is included with the MCP tool request, enabling the MCP server to enforce subject, actor, and authorization context before dispatching the tool.	View
agent-identity-vault-course-agent-sample	Extends the direct tool pattern to protect tool credentials. Premium course enrollment requires an API key that is never stored in the agent or source code. HashiCorp Vault releases the key only after validating an IBM Verify delegated access token against the Agent Registry, scope, and RAR policy.	View