IAMIDENTITY.AI

Identity-First Security
for the Agentic AI Era

Zero trust architecture, workload identity, continuous authorization, and delegated access — written from the trenches of enterprise IAM.

Articles

Shared Signals MCP IBM Verify

Securing the MCP, Part Two: Shared Signals and the Kill Switch

Three denials. One signed event. Tenant-wide revoke. Move revocation into the MCP server using SSF/CAEP and let IBM Verify enforce it across every federated app. The same pipeline runs any API call you wire to it.

MCP IBM Verify HashiCorp Vault

The MCP Server Is Your Security Perimeter

An identity pattern for MCP tool calls you can stand up in 30 minutes — no AWS account required. Verify decides, Vault mints, and the credential lives five minutes. A hands-on cookbook and the why behind it.

AWS Bedrock AgentCore IBM Verify Healthcare

AgentCore Runs the Loop. Verify Runs the Trust.

A healthcare AI agent on AWS Bedrock AgentCore Runtime. The loop runs in AWS. Every credential, every authorization decision, every short-lived database lease still belongs to IBM Verify, HashiCorp Vault, and SPIFFE.

Vercel Breach OAuth Supply Chain IBM Verify + Vault Agentic Runtime

When the AI Integration Becomes the Attack Vector — The Vercel Breach Was an Architecture Problem

Three hops, no zero-day. The Vercel breach started in an AI tool, pivoted through OAuth, and ended in plaintext environment variables. Every failure mode in the chain has a direct architectural counter.

Agentic AI RFC 8693 IBM Verify Token Exchange

One Chat, Three Specialists, Zero Standing Privileges

A Claude hosted supervisor agent fans one user message out across IBM Verify, Microsoft Entra ID, and a banking system. Every hop carries its own delegation chain. The credential that ran the SQL did not exist sixty seconds ago.

RFC 9396 IBM Verify HashiCorp Vault Dynamic Credentials

Binding RAR to Vault: We Built the Plugin

We bound IBM Verify's Rich Authorization Requests to HashiCorp Vault dynamic credentials. The database role that ran your transaction did not exist before you tapped Approve. It will not exist after.

Claude-Powered Agent IBM Verify Token Exchange SPIFFE

IBM Identity Agent: Securing a Claude-Powered Agent with IBM Verify

The model runs on Anthropic's infrastructure. The secrets live in your Vault. The authorization decision happens in IBM Verify. A working hosted-agent security pattern with Token Exchange, RAR, SPIFFE, and CAEP — deployed today.

Agentic AI IBM Verify Entra ID Identity Lab

Build Your Agentic AI Identity Lab — For Free

A full identity ecosystem — Entra ID, IBM Verify federation, agent identity, Token Exchange, SPIFFE, RAR — built on a free developer tenant. Here's how.

RSAC 2026 Agentic AI IBM Verify

"So You Have a Dashboard?" — RSAC 2026 Field Notes

RSAC 2026 was full of flashy screens and marketing fluff. We showed a working agent security runtime. Most of the floor showed a dashboard. Here's what I actually saw at Moscone.

Agentic AI Token Exchange IBM Verify RAR

Can You Secure Agentic AI Without Buying Another Product?

A Deputy CISO asked the question every identity team should be asking: can we secure our AI agents with what we already have? The answer depends on your identity ecosystem — and most aren't ready.

Agentic AI Identity Zero Trust

The Weakest Link: Agentic AI Agents

The weakest link used to be humans. Now it's agents. We spent decades building identity governance around people — then handed autonomous systems the exact same anti-patterns and walked away.

AWS EC2 HashiCorp Vault IBM Verify SPIFFE

Secretless by Design: Zero-Trust Agentic AI on AWS EC2

How I deployed a LangChain-powered agentic AI with no secrets in the code, no Vault tokens on disk, cryptographic identity via SPIFFE, continuous session evaluation with CAEP, and delegated authorization through Token Exchange and RAR.

The Author

Robert Graham

Global Product Architect

IBM Verify · IBM Security

Robert is a Global Product Architect at IBM Verify with over a decade at IBM spanning cloud security consulting, identity architecture, and technical sales engineering. He specializes in zero-trust architecture, workload identity, and enterprise IAM — with deep hands-on expertise across the protocols that underpin modern identity: OAuth 2.0, OIDC, SAML, FIDO2, SCIM, SPIFFE, CAEP/SSF, Token Exchange, and Rich Authorization Requests.

His current work focuses on the intersection of agentic AI and identity security — building and documenting architectures that bring mutual zero trust, secretless infrastructure, and continuous authorization to autonomous AI systems. He architects solutions using IBM Verify, HashiCorp Vault, SPIFFE/SPIRE, and the Shared Signals Framework, and writes about what he builds at iamidentity.ai.

Zero Trust IBM Verify HashiCorp Vault SPIFFE / SPIRE CAEP / SSF OAuth 2.0 OIDC FIDO2 SAML SCIM LangChain Agentic AI Token Exchange RAR

LinkedIn → Medium → iamidentity.ai →