We are excited to announce the first release of RSA Agent ID, a platform for discovering, securing, and governing the AI agents and Non-Human Identities (NHIs) that act inside your environment — from autonomous agents and AI copilots to the MCP servers, service accounts, and API keys that support them.
This release includes two capabilities:
- Discover: finds and risk-profiles every agent and NHI across your identity providers, cloud platforms, and endpoints.
- Secure: a policy-enforcing gateway that authenticates, authorizes, and audits every action an agent takes.
You can license either capability or both, and manage them from a single cloud-hosted admin console.
Discover: Find Every Agent and NHI
Shadow Asset Discovery
Surfaces agents, AI processes, and MCP servers that have no registry entry, each with its detection type, the source that found it, and the evidence behind its risk. A summary shows how many were first detected in the last seven days.
Unified Agents, Resources, and Owners Registry
Tracks every discovered or registered agent, NHI, and resource in one registry, with its status, owner, and asset risk. Asset risk is the highest severity among an asset's open findings.
Agent Profiles
Groups agent instances under a single profile with a shared type (such as a Platform, Local, IDE, CLI, SaaS, or Gateway agent), owner, and sanction status. Matched instances take the profile's sanction status and keep their own owner and review decisions. You can mark a profile for review or block it.
Blast Radius View
Displays a graph of what an agent or NHI can reach if it is compromised: the owner, the identity it runs as, its permissions, and the target APIs, resources, or devices. Select a node to view its relationship and permissions, and expand grouped nodes to see every agent that shares an identity.
Owner Accountability and Correlation
Correlates each agent and NHI to the person accountable for it, using identity-provider and endpoint data. Inferred owners appear in the graph, and you can assign, search, or notify owners from the registry.
Posture Findings and Detections
Evaluates identities against posture checks. Each finding displays its contributing factor, links to evidence, and recommends an action.
Discover Overview
Summarizes connector health, the identity count by type, prioritized investigations, the risk landscape, unmanaged AI activity, and ownership gaps.
Connector Catalog with Guided Setup
Provides ready-to-use connectors (see Supported Identity Providers and Data Sources) that sync on a schedule. Guided setup generates the required grant, such as an admin consent link, an AWS role template, or an Okta key, for you to run in your own console. Saving a connection starts a discovery run.
Secure: Control Agent Access Through the Gateway
Flexible Gateway Deployment
Registers a gateway in the console. Each tenant includes an RSA-hosted gateway runtime. You can also deploy gateway runtimes in your own Kubernetes environment in Azure or AWS, or in an on-premises environment.
Resource Registration for MCP Servers and REST APIs
A guided wizard registers any MCP server or OpenAPI/Swagger REST API behind a gateway. The gateway discovers each tool and suggests a Low, Medium, or High risk level that you can adjust.
Secured Agents
Lists agents registered with Agent ID Secure, with their publish status, asset risk, owner, and who enforces runtime control. You can block an agent. Only agents with Source: Agent ID Secure can be blocked.
Roles and Scopes
Grants tool and action access through roles mapped to identity-provider groups, so access follows group membership.
Policy Engine with Audit and Enforcing Modes
Evaluates policies at the gateway on every agent call, scoped tenant-wide or to specific gateways, and compiles them to Cedar. Run a policy in Audit mode before switching it to Enforcing; a forbid rule always overrides a permit. Policy templates, system defaults, and a summary of decision outcomes are included.
Human-in-the-Loop Approvals
Routes sensitive actions to a human reviewer by inline approval in MCP clients that support Elicitation, out-of-band backchannel approval (CIBA), or an email redirect. The Approval requests page tracks each request.
Controlled Publishing
Resources, roles, policies, agents, and IdP bindings take effect on a gateway only after you publish them. Publish management shows pending changes, publish history, and who published; a publish status column appears on every governed table.
Administer: Run the Tenant with Confidence
Zero-touch Bootstrap
Each tenant includes a pre-created break-glass administrator account. Sign in, change the password, and accept the license agreement. Then connect your identity provider, such as Microsoft Entra ID or Okta, to onboard additional administrators.
Administrator Roster and Group Mapping
Lists everyone in your administrator groups, signed in or not, with the IdP group that granted access, role, last sign-in, and status.
General Settings and Email Templates
General settings let you edit the organization name and view RSA-managed system information. Email templates cover owner assignment and human approval requests, with a preview.
Audit Log
Records administrative activity, connector syncs, and every agent tool call in one filterable, exportable log. Records are tamper-evident and isolated by tenant.
Built for How You Deploy
Control-plane/Data-plane Architecture
You configure RSA Agent ID once in the cloud-hosted control plane. Each gateway runtime, which acts as the data plane, enforces that configuration locally without direct internet exposure.
Resilient Gateway Runtime
Ships as container images and a Helm chart. Run two or more replicas with a managed PostgreSQL and Redis for high availability.
Supported Identity Providers and Data Sources
RSA Agent ID connects to the identity providers your administrators use to sign in, and to the sources Discovery reads to find agents and NHIs.
Administrator Identity Providers
| Identity provider | Supported protocol |
|---|---|
| Microsoft Entra ID | OIDC |
| Okta | OIDC |
| Keycloak | OIDC |
| PingOne | OIDC |
Discover Data Sources
| Source | Type | What it contributes |
|---|---|---|
| Microsoft Entra ID | Identity provider | Users, groups, applications, and service principals. |
| Okta | Identity provider | OAuth apps, API tokens, and service integrations with their scopes and admin roles, and AI agents (Okta for AI Agents). |
| Microsoft Defender | Endpoint/EDR | Agents, devices, and endpoint telemetry. |
| CrowdStrike Falcon Next-Gen SIEM | Endpoint/EDR | Agents and processes on managed endpoints. |
| AWS | Cloud | IAM users, roles, access keys, and Bedrock agents with their execution roles. |
| MCP Gateway | Gateway | Agent ID Secure gateway and agents registered in Agent ID Secure, with their owners. |
RSA Agent ID Plans
RSA Agent ID is licensed in modules so you can start with what you need and expand later:
| Plan | What’s included |
|---|---|
| Discover | Risk (Shadow assets, Posture findings), Registry (Agents/NHI, Agent profiles, Resources, Owners), Monitor (Audit), Settings (Connectors, Integrations, Organization). |
| Secure | Registry (Agents, Resources, Owners), Monitor (Approval requests, Audit), Secure (Gateways, Roles and scopes, Policies), Settings (Connectors, Integrations, Organization). |
| Discover + Secure | Both capability sets in a single tenant. |
Product Documentation and Support
Visit RSA Community to access the latest product documentation, answers to common questions, solutions to known problems, community discussions, and case management.
Related Articles
RSA November 2026 Release Announcements 13Number of Views RSA February 2026 Release Announcements 47Number of Views RSA March 2026 Release Announcements 35Number of Views RSA November 2024 Release Announcements 174Number of Views RSA July 2026 Release Announcements 34Number of Views