Join our Newsletter — 33% off our NHI Course
Home› Guides› How to centralise root access in AWS Organizations…
Guide Cloud & Workload Identity

How to centralise root access in AWS Organizations and delete member account root credentials: Enable, Audit, Delete and Recover Safely

PDF for newsletter subscribers · free to read online

← All guides
By NHI Mgmt Group, edited by Lalit Choda Updated 20 November 2026 14 min read
On this page

Centralised root access in AWS Organizations lets the management account, or a delegated administrator for IAM, remove long-term root user credentials from member accounts and run the few root-only tasks through short, scoped privileged sessions instead. This how-to covers the whole procedure: prerequisites and permissions, enabling the two capabilities, auditing which member accounts still hold root credentials, deleting them, handling the tasks that still need root, and recovering a root user deliberately when you have to.

Every member account in an organisation starts life with a root user that has complete access to all services and resources in that account. Each of those root users is a long-lived credential: a password, possibly access keys, possibly signing certificates, and an MFA device that someone has to keep safe. Multiply that by hundreds of accounts and the root estate becomes one of the largest and least visible sets of privileged credentials you own. Centralised root access is how AWS lets you shrink it to almost nothing.

This is an independent guide from NHI Mgmt Group, based on AWS's public documentation as of November 2026. It is not sponsored or reviewed by AWS. Product features change, so confirm settings against AWS's current documentation.

Key takeaways

  • Centralised root access has two capabilities: Root credentials management, which lets you delete member root credentials, and Privileged root actions in member accounts, which lets you run root-only tasks without signing in as root.
  • Deleting root credentials removes the root user password, access keys and signing certificates, and deactivates MFA for the member account.
  • New accounts created in AWS Organizations have no root user credentials by default once the feature is in place.
  • Privileged tasks run through sts:AssumeRoot with one of five AWS managed task policies, called against a Regional STS endpoint, not the global endpoint.
  • Root credentials cannot be used to call sts:AssumeRoot; it must be an IAM user or role in the management or delegated administrator account with the permission explicitly granted.
  • If you need a member root user back, the Allow password recovery task restores the path, and AWS recommends deleting the credentials again once the task is done.

What centralised root access changes

AWS's root user documentation explains that, after you centralise root access for member accounts, you can choose to delete root user credentials from member accounts. You can remove the root user password, access keys and signing certificates, and deactivate MFA. Member accounts then cannot sign in to their root user or perform password recovery for their root user unless account recovery is enabled.

That changes the threat model for each member account. A root user with no password, no keys and no MFA device has nothing to phish, nothing to leak in a repository, and no recovery path that an attacker who controls the root email inbox can use without first being granted one. The account still exists and still has a root user, but the credential that made it dangerous is gone.

Some tasks have always required root. With centralised root access, AWS documents that the management account or delegated administrator for IAM can perform certain of them through short-term privileged sessions. These sessions provide temporary credentials scoped to specific privileged actions, without root user sign-in on the member account. The documented actions are deleting a misconfigured Amazon S3 bucket policy, deleting a misconfigured Amazon SQS queue policy, deleting root user credentials for a member account, and re-enabling root user credentials for a member account.

AWS also notes that some tasks can still only be performed when you sign in as the root user of an account. For those, you use the recovery path described later rather than keeping standing credentials on every account.

In framework terms this is a direct application of NIST SP 800-53 Rev. 5 AC-6 Least Privilege and AC-2 Account Management, and it removes one route for MITRE ATT&CK: Valid Accounts: Cloud Accounts (T1078.004).

Prerequisites and permissions

The AWS procedure for centralising root access has one account requirement: you must manage your AWS accounts in AWS Organizations. Plan the following before you begin.

Who will operate the feature

Decide whether root access will be run from the management account or from a delegated administrator for IAM. AWS recommends that a delegated administrator be an account intended for security or management purposes. Using a delegated security account keeps day-to-day operation out of the management account, which should itself have as few human sign-ins as possible. Our AWS Organizations and SCPs Guide covers how to structure those accounts.

Permissions to enable the feature

AWS lists the permissions needed to enable centralised root access:

  • iam:EnableOrganizationsRootCredentialsManagement
  • iam:EnableOrganizationsRootSessions
  • iam:ListOrganizationsFeatures
  • organizations:EnableAwsServiceAccess
  • organizations:ListAccountsForParent
  • organizations:RegisterDelegatedAdministrator

For full console functionality AWS also recommends read permissions such as organizations:DescribeAccount, organizations:DescribeOrganization, organizations:ListAWSServiceAccessForOrganization, organizations:ListDelegatedAdministrators, organizations:ListOrganizationalUnitsForParent, organizations:ListParents and organizations:ListTagsForResource.

Permissions to audit root credentials

Auditing member root credentials needs iam:GetAccessKeyLastUsed, iam:GetAccountSummary, iam:GetLoginProfile, iam:GetUser, iam:ListAccessKeys, iam:ListMFADevices, iam:ListSigningCertificates and sts:AssumeRoot. AWS notes that you can use the IAMAuditRootUserCredentials AWS managed policy, or any policy with access to iam:GetAccountSummary, to audit the root credential status of a member account.

Grant these to a small, named IAM role, not to a broad administrator role. Whoever holds sts:AssumeRoot can act on member root users, so that role is now one of the most privileged identities in your organisation.

Step 1: Enable trusted access and the two capabilities

Console

  1. Sign in to the AWS Management Console and open the IAM console.
  2. In the navigation pane, choose Root access management, then select Enable. If you see Root access management is disabled, AWS directs you to enable trusted access for AWS Identity and Access Management in AWS Organizations first.
  3. In the Capabilities to enable section, choose Root credentials management to allow the management account and delegated administrator to delete root user credentials for member accounts.
  4. Choose Privileged root actions in member accounts to allow the management account and delegated administrator to perform certain tasks that require root user credentials.
  5. Optionally, enter the account ID of the Delegated administrator.
  6. Choose Enable.

Enable both capabilities together. AWS states that you must enable privileged root actions in member accounts to allow member accounts to recover their root user credentials after they have been deleted. Turning on deletion without the recovery capability leaves you without the documented way back.

AWS CLI

The equivalent CLI sequence documented by AWS is to enable trusted access for IAM in Organizations if it is not already enabled, then enable each capability, then optionally register the delegated administrator:

  • aws organizations enable-aws-service-access --service-principal iam.amazonaws.com
  • aws iam enable-organizations-root-credentials-management
  • aws iam enable-organizations-root-sessions
  • aws organizations register-delegated-administrator --service-principal iam.amazonaws.com --account-id 111111111111

The matching API operations are EnableAWSServiceAccess and RegisterDelegatedAdministrator in the Organizations API, and EnableOrganizationsRootCredentialsManagement and EnableOrganizationsRootSessions in the IAM API. If you manage your organisation as code, put these in the same pipeline that owns your other organisation-level settings so the state is reviewable.

Step 2: Audit which member accounts hold root credentials

Before deleting anything, build an inventory. For each member account you need to know whether the root user has a password, access keys, signing certificates or an MFA device, and whether any access key has been used recently.

AWS documents the operations that check root credential status inside a privileged session: get-user, get-login-profile, list-access-keys, list-signing-certificates, list-mfa-devices and get-access-key-last-used on the CLI, or GetUser, GetLoginProfile, ListAccessKeys, ListSigningCertificates, ListMFADevices and GetAccessKeyLastUsed through the API. Run them in a session started with the IAMAuditRootUserCredentials task policy.

Record the results per account and sort them into three groups:

  • No root credentials: nothing to do beyond monitoring.
  • Password and MFA only: a standard deletion candidate. Confirm with the account owner that no one relies on root sign-in.
  • Root access keys present: treat as a priority. Use get-access-key-last-used to see whether the key is in use. A root access key that is still being used means some process or person is authenticating as root, and that dependency must be moved to an IAM role before deletion or it will break.

Root access keys are the clearest non-human identity risk in this inventory. They are long-lived, unscoped and usually forgotten. Our AWS Security Guide covers moving workloads from long-term keys to roles.

Step 3: Delete member root credentials

Console

  1. In the IAM console, choose Root access management.
  2. Select the member account from the list and choose Take privileged action.
  3. Choose Delete root credentials.

AWS documents that deleting root user credentials removes the root user password, access keys and signing certificates, and deactivates MFA for the member account.

CLI and API

From the CLI, start a privileged session against the member account with the deletion task policy. AWS's example is:

aws sts assume-root --target-principal 111122223333 --task-policy-arn arn=arn:aws:iam::aws:policy/root-task/IAMDeleteRootUserCredentials --duration-seconds 900

Use the returned AccessKeyId, SecretAccessKey and SessionToken for the following calls. AWS lists the deletion operations as delete-login-profile, delete-access-key, delete-signing-certificate and deactivate-mfa-device, or DeleteLoginProfile, DeleteAccessKey, DeleteSigningCertificate and DeactivateMfaDevice through the API.

AWS notes that the global endpoint is not supported for sts:AssumeRoot, so you must send the request to a Regional AWS STS endpoint. Its API example uses sts.us-east-2.amazonaws.com. Pin the Region explicitly in any script so it does not fall back to a default that resolves to the global endpoint.

Rolling it out across the organisation

Work in waves: sandbox and development accounts first, then shared services, then production. For each wave, re-run the audit afterwards to confirm the account now reports no root credentials, and keep the before and after records as evidence for auditors. Hold back any account where the audit found an active root access key until its owner has moved the dependency.

Step 4: Run root-only tasks through privileged sessions

AWS documents that to use centralised root access you must sign in through the management account or a delegated administrator account as an IAM user or role with the sts:AssumeRoot permission explicitly granted. You cannot use root user credentials to call sts:AssumeRoot.

Each session must name a task policy. AWS lists five AWS managed task policies:

Task policyWhat it is for
IAMAuditRootUserCredentialsReading root credential status for a member account
IAMCreateRootUserPasswordAllowing password recovery for a member root user
IAMDeleteRootUserCredentialsDeleting member root user credentials
S3UnlockBucketPolicyDeleting a misconfigured S3 bucket policy that denies all principals
SQSUnlockQueuePolicyDeleting a misconfigured SQS queue policy that denies all principals

The S3 and SQS tasks solve a familiar problem: a resource policy that denies every principal, including administrators, which previously needed root sign-in to fix. In the console you choose the member account, Take privileged action, then either Delete bucket policy after browsing to the bucket, or Delete SQS policy after entering the queue name. You then correct the policy in the S3 or SQS console.

To limit what a given caller can do during a privileged session, AWS documents the STS condition key sts:TaskPolicyArn. Use it in the identity policy that grants sts:AssumeRoot so that, for example, an operations role can only use the S3 and SQS unlock policies while only a security role can use the credential deletion and password creation policies. That separation stops one compromised role from being able to both restore a root password and use it.

AWS's examples use a session duration of 900 seconds. Keep sessions short and scripted: one task, one session.

Step 5: Recover a member root user when you genuinely need it

Some tasks still require root sign-in. When one comes up, AWS documents the Allow password recovery privileged task, performed by the management account or delegated administrator. After that action, the person with access to the root user email inbox for the member account can reset the root user password and sign in to the member account root user. AWS notes that this option is only available when the member account has no root user credentials.

Through the CLI, the operations are get-login-profile and create-login-profile in a session started with the IAMCreateRootUserPassword task policy.

Treat every recovery as a change with a ticket, an approver and an end time:

  1. Record why root sign-in is needed and confirm the task really is root-only.
  2. Run Allow password recovery from the delegated administrator.
  3. Have the holder of the root email inbox reset the password, register MFA, and perform the task.
  4. Run Delete root credentials again as soon as the task is complete, which AWS recommends.
  5. Re-run the audit and attach the result to the ticket.

The control that remains in this model is the root email inbox. Whoever controls it can complete recovery once it has been allowed, so the inbox must be a managed, monitored mailbox, not an individual's address. For the separate case where a standalone account's root MFA or email is lost, see How to recover AWS root access when the MFA device or email is lost.

Monitoring and governance after rollout

Once member root credentials are gone, the interesting events are the ones that bring them back or use the privileged path. Build the following into your monitoring and governance:

  • Privileged session use: log and review every AssumeRoot call, including which task policy was used and against which member account. Our AWS CloudTrail Guide covers organisation trails and alerting.
  • Recovery events: alert on every use of the password creation task policy and confirm it maps to an approved change.
  • Drift: re-run the root credential audit on a schedule. Any member account that reports root credentials outside an approved recovery window is an incident.
  • Role ownership: the roles with sts:AssumeRoot need named owners, MFA-protected sign-in for humans, and periodic access reviews.
  • Management account: AWS documents centralised root access for member accounts, so plan the protection of the management account's own root user separately. The AWS Root Account Compromise Playbook covers response if a root user is compromised.

Common failure patterns

  • Deletion without the recovery capability: enabling root credentials management but not privileged root actions removes the documented route to recover a member root user.
  • Deleting before auditing: removing an active root access key breaks whatever depended on it, and the incident that follows makes teams reluctant to finish the rollout.
  • Broad sts:AssumeRoot grants: giving the permission to a general administrator role, without sts:TaskPolicyArn conditions, turns one compromised role into control of every member root user.
  • Global STS endpoint in scripts: automation that calls the global endpoint fails, because AWS does not support it for sts:AssumeRoot.
  • Recovery left open: allowing password recovery for a one-off task and never deleting the credentials again recreates the standing root credential the project set out to remove.
  • Unmanaged root email inboxes: personal or unmonitored addresses on member accounts leave the recovery path in the hands of whoever can read the mail.
  • Running everything from the management account: skipping a delegated administrator puts routine privileged work in the account that should be touched least.

Practitioner checklist

  • Confirm all accounts are in AWS Organizations and choose a security account as the delegated administrator for IAM.
  • Enable trusted access for IAM, then enable both Root credentials management and Privileged root actions in member accounts.
  • Create a dedicated audit role and run the root credential audit across every member account with IAMAuditRootUserCredentials.
  • Move any workload still using a root access key onto an IAM role before deletion.
  • Delete member root credentials in waves and re-audit after each wave.
  • Restrict sts:AssumeRoot with sts:TaskPolicyArn so deletion, recovery and resource unlock tasks sit with different roles.
  • Pin scripts to a Regional STS endpoint and keep session durations short.
  • Run recovery only through a ticketed change and delete the credentials again afterwards.
  • Move member root email addresses to managed, monitored mailboxes.
  • Alert on every privileged session and every password recovery event, and re-audit for drift on a schedule.

Standards and references

Related NHI Mgmt Group resources: AWS Root Account Compromise Playbook · AWS Organizations and SCPs Guide · How to recover AWS root access when the MFA device or email is lost · AWS Security Guide

PDF for subscribers

Take this guide with you

Every section, table and checklist in one PDF, with links to the sources and related guides. Subscribe to our weekly NHI & AI Identity Journal and we’ll email you a download link.

Free weekly newsletter

Subscribe to the NHI & AI Identity Journal

The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.

Bonus 33% off our NHI Course when you subscribe.

NHIMG Editorial Note
Researched and written by the NHI Mgmt Group editorial team using AI tools, under the editorial direction of Lalit Choda. Facts are drawn from the sources listed on this page and checked by our automated quality process. Last updated 20 November 2026. Spotted an error? Use our contact page.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM and Agentic AI security. nhimg.org

PDF Downloads for Newsletter Subscribers

Get the guide

Enter your details and we’ll email you a download link.

The full guide is free to read online. The PDF is a thank you for subscribing.

Fill in all fields marked * and tick the PDF box to continue.