Skip to content
Rootways CIM Pro flaw CVE-2026-130583 lets attackers read Magento 2 store keys
Black Shard

Loading

    Essential Eight

    Multi-factor authentication, from sensitive services to every data repository.

    The strategy says who must prove their identity twice and how strong the second proof must be. The first level covers the services that hold sensitive data; the second brings in every user of every system with phishing-resistant factors; the third adds the data repositories.

    Not published

    What the strategy requires

    At Maturity Level One multi-factor authentication protects users of the organisation's own online services that hold its sensitive data, users of third-party services that hold sensitive data, users of other third-party services where it is available, and both staff and customers of online customer services that hold sensitive customer data.

    The factors must be something users have and something they know, or something they have unlocked by something they know or are. Maturity Level Two extends the control to privileged and unprivileged users of systems, requires phishing-resistant factors for online services and systems, gives customers a phishing-resistant option, and centrally logs successful and failed authentication events.

    Maturity Level Three adds users of data repositories, makes customer authentication phishing-resistant outright, and extends log analysis from internet-facing servers to every server and workstation.

    What an assessor looks for

    The assessor watches a sign-in. For each service in scope they ask a user to log in and note which factor is presented and whether the service can be reached without it from an unmanaged device. They read the identity provider's policy for exclusions: the service account, the shared mailbox, the executive who asked to skip the prompt.

    From the second level they test the factor's resistance to phishing. A one-time code typed into a web page is not phishing-resistant; a security key or a platform authenticator bound to the site is. They look for the central log of authentication events and for someone reading it.

    Common gaps

    Coverage is the first gap: multi-factor authentication is on for the mail service and off for the remote desktop gateway, the finance platform and the backup console. Strength is the second: codes by text message and push notifications that a tired user approves, which fall short of the second level's phishing-resistant requirement for staff.

    Third-party services are the third gap, where the vendor offers the factor and the organisation never turned it on. Customer-facing services are missed because the organisation thinks of the strategy as an internal control; from the first level, customers of a service that holds their sensitive data must authenticate with a second factor.

    How to meet it in Microsoft 365 and Azure

    Microsoft Entra ID is the identity provider for the first two levels. Entra ID Conditional Access policies require authentication strength per application and per user group, with phishing-resistant strength for administrators and then for everyone, met by FIDO2 security keys, Windows Hello for Business or passkeys in Microsoft Authenticator.

    Entra ID Identity Protection raises the bar on risky sign-ins, and third-party services are brought under the same policy by federating them to Entra ID. Successful and failed authentication events are Entra ID sign-in logs, exported to a Log Analytics workspace and read in Microsoft Sentinel for the central logging and analysis the second level requires.

    Data repositories at the third level, including SharePoint Online, OneDrive and Azure storage, take the same Conditional Access strength because access to them passes through Entra ID.

    Where it sits on the ladder

    The first level is the sensitive services and the customer services; the second is every user of every system with phishing-resistant factors and central logs; the third is the data repositories and customers on phishing-resistant factors too.

    Restrict administrative privileges is the companion strategy: the privileged users it separates are the ones the second level makes authenticate with the strongest factors first.

    Dates

    When it was written, and checked.

    .

    Questions, answered.

    Is a code sent by text message acceptable?

    At Maturity Level One, yes, as something the user has. From Maturity Level Two the factors used by staff for online services and systems must be phishing-resistant, which a code typed into a page is not.

    What makes a factor phishing-resistant?

    The proof is bound to the real site or device so it cannot be relayed through a fake one: a FIDO2 security key, a platform authenticator such as Windows Hello for Business, or a passkey. A push approval or a typed code can be relayed.

    Do customers of our portal need multi-factor authentication?

    If the portal holds their sensitive data, yes, from Maturity Level One. At the second level they must have a phishing-resistant option, and at the third level the factor used must be phishing-resistant.

    Does a password manager count as a second factor?

    No. A stored password is still something the user knows. The second factor is something the user has, or has that is unlocked by something they know or are.

    Provenance

    Sources, and when they were read.

    Every source was read 31 Oct 2026.

    Being measured against Essential Eight? Scope the work.