Consent Phishing: When a Fake App Gets Into Microsoft 365

Most phishing tries to steal a password. Consent phishing skips the password entirely. The attacker sends a link to a real Microsoft sign-in page, and after the user signs in, Microsoft asks them to give an app permission to “read your mail” or “access your files.” One click on Accept, and the attacker’s app can read that mailbox or OneDrive through Microsoft’s own systems, with no password to steal and no MFA prompt to beat. The fix is a setting most small businesses have never looked at: who in your organization is allowed to approve apps.

How a consent phishing attack works

  1. The lure. An email invites the user to open a shared document, sign an agreement, or try a new productivity tool. The link leads to a genuine Microsoft sign-in page, so nothing looks wrong.
  2. The permission prompt. After sign-in, Microsoft shows the app’s name and the permissions it wants: read your mail, read and write your files, keep access to data you have given it access to. The app usually has a convincing name and logo.
  3. The grant. When the user accepts, Microsoft issues the app access to act on that user’s behalf. The attacker can now read mail, download files, or send messages as the user, from their own servers, at any hour.
  4. The persistence. The access belongs to the app, not to the password. Resetting the user’s password or re-registering their MFA does not reliably remove it; the grant itself has to be revoked.

Why it gets past the usual defenses

MFA protects the sign-in, and the user really did sign in. The email points to a legitimate Microsoft address, so link scanners have less to flag. And the app asks for the same kinds of permissions that legitimate tools ask for every day: scheduling apps, e-signature services, CRM add-ins, AI assistants. A user who approves apps routinely has no reason to hesitate. That is why consent phishing works against businesses that have already done the basics, including phishing-resistant MFA.

The setting that stops most of it

Microsoft Entra ID decides whether employees can approve apps on their own. In the Microsoft Entra admin center, under Enterprise applications > Consent and permissions, the main choices are:

  • Allow user consent for apps. Any user can give any app access to their data. This is the setting attackers count on.
  • Allow user consent for apps from verified publishers, for selected permissions. Users can approve only apps whose publisher Microsoft has verified, and only for low-risk permissions you choose. Microsoft recommends this option.
  • Do not allow user consent. Only an administrator can approve apps.

For most small businesses, the verified-publisher option or full admin approval is the right call. Pair either one with the admin consent workflow, so an employee who needs an app can request it instead of hitting a dead end. The request goes to an administrator, who checks the app and approves it once for everyone who needs it. Microsoft Secure Score also lists restricting user consent as a recommended action; see what to fix first in Secure Score.

Check what has already been approved

Changing the setting stops new grants, not old ones. Review the apps your users have already approved under Enterprise applications, and look closely at:

  • Apps nobody recognizes, or with generic names like “Document Viewer” or “Secure Files.”
  • Apps with broad permissions, such as reading all mail or all files.
  • Apps approved by a single user, especially someone in finance, payroll, or leadership.
  • Apps from publishers that are not verified.

Plenty of what you find will be legitimate: a calendar tool, a scanner app, an e-signature service. The goal is to know what each app is, who uses it, and whether it still needs the access it has. Remove anything nobody can explain.

If you find a malicious app

  1. Revoke the app’s permissions, then disable or delete it under Enterprise applications.
  2. Revoke the affected users’ sessions so any access already issued stops working.
  3. Find out what the app touched. Check the audit log for mail and file activity, and check the mailbox for forwarding rules or sent messages the user did not write.
  4. Treat it as a security incident. Follow your incident response plan, and work out whether customer, employee, or financial data was exposed.

Train for the prompt, not just the link

Security awareness training usually teaches people to check links and sender addresses. Add one more rule: a screen asking you to give an app access to your mail or files is a decision, not a formality. If you don’t recognize the app, or you didn’t go looking for it, click Cancel and ask IT. The SLAM method still applies to the email that started it, and if you suspect a mailbox has already been accessed, see how to tell if your email is hacked.

Lock down app consent

Equal Tech Solutions configures Entra ID consent settings, reviews the app permissions already granted in your tenant, and watches for risky new grants as part of our cybersecurity services and managed IT for businesses across Cleveland, Chattanooga, and the Southeast. Contact us to find out which apps can read your company’s email today.