Consent phishing: the app that asks for your permissions
A fake cloud app gets access to your mailbox or files with one click on Accept, with no stolen password involved.
On this page
In 30 seconds
- Where it happens
- In the company's cloud, when a third-party application asks for permission to access the user's information and to sign in on their behalf in other cloud applications[2].
- What the attacker wants
- An access token. Once granted, the application can keep potentially long-lasting access to the account's functions[3].
What it is
A malicious application can get access to an employee's mailbox or files without knowing their password. The employee authorizes it, on a consent screen, by clicking Accept. Microsoft calls these attacks "consent phishing": they aim to push users into granting permissions to malicious OAuth applications so that those applications can reach their legitimate cloud services, such as email servers, file storage and management APIs[1].
OAuth is an open standard for authentication and authorization by token. It lets a third-party service use information from a user's account without exposing their password: the application receives an access token that authorizes the sharing of specific information[2].
Straight answers
Does the attacker need my password?
No. According to MITRE ATT&CK, an OAuth access token lets a third-party application interact with resources that hold the user's data, in the way the application requested, without obtaining the user's credentials[3]. The attacker can sometimes also steal a refresh token, which gives them new access tokens without asking the user[3].
In the December 2022 campaign analyzed by Proofpoint, the permissions included "offline access": access no longer needs any action from the user after consent, and the refresh token was in most cases valid for more than a year[5]. In an attack Microsoft described in 2022, where the application had been created from compromised administrator accounts, the actor had added their own credentials to the application and could use them even if the administrator changed their password[1].
Is the "verified publisher" badge a guarantee?
No. According to Proofpoint, attackers met Microsoft's requirements by abusing the verified publisher status, and users and organizations should not trust an OAuth application on that status alone[5]. Microsoft, for its part, recommends allowing user consent only for applications published by a verified publisher[4]. Proofpoint also advises using Microsoft's "verified publisher" policy[6].
How can I spot a suspicious request?
MITRE ATT&CK recommends training users not to authorize third-party applications they do not recognize. The redirect URL deserves a look: if it is a misspelled or convoluted string of words tied to an expected service, the site is probably trying to impersonate a legitimate service. Offline access and the ability to read email should raise more suspicion[3]. Microsoft adds that an application should only ask for permissions related to its purpose, and gives the example of a clock application with full access to all mailboxes[2].
What can administrators restrict?
According to MITRE ATT&CK, they can block user consent to OAuth applications, which forces administrator consent for every request, and block users from registering applications[3]. Microsoft describes a built-in policy, microsoft-user-default-low, that limits consent to applications from verified publishers and applications registered in the tenant, for permissions classified as low impact only[4]. Proofpoint also recommends limiting application creation by non-administrator accounts[6].
Inside the attacker's head
According to MITRE ATT&CK, the attacker builds an application designed to receive access to resources through the target's OAuth token. They register it with the authorization server, for example Microsoft's identity platform, then send the target a targeted phishing link to induce them to grant access[3].
They can polish the appearance. In the December 2022 campaign, the displayed publisher name imitated that of an existing legitimate publisher, while the verified publisher name stayed hidden and different, for example "Acme Holdings LLC" instead of "Acme LLC"[5]. The attackers added links to terms of use and a policy statement that pointed to the website of the organization being imitated, and Proofpoint assumes they added credibility[5]. According to Proofpoint, the possible consequences of this campaign are data exfiltration, business email compromise (BEC) and mailbox abuse[5].
They can also go through an account that is already compromised. Proofpoint names MACT the technique of creating new applications in hijacked environments to launch further attacks and establish lasting persistence[6].
What you see on screen
In the December 2022 campaign, two applications were called "Single Sign On (SSO)" and the third "Meeting". They used an old version of the Zoom icon and redirect URLs that looked like Zoom[5].
Two of these applications asked for these permissions: User.Read, email, offline_access, profile, openid, Mail.Read and MailboxSettings.Read. The third added Calendars.read, Onlinemeetings.read and Mail.send[5]. According to Microsoft documentation cited by Proofpoint, a blue verification badge appears in the consent prompt when the publisher has been verified[5].
Microsoft notes that users often click accept without closely examining the details of the prompt, including the permissions[2]. By default, a user can authorize an application to access their mailbox, but not to read and write all files across the organization without limit[4].
Under the hoodIT teams
MITRE ATT&CK cites Microsoft's Authorization Code Grant flow as a common example of how the token is issued[3]. According to Proofpoint, about 16% of the organizations it monitored in 2023 experienced at least one incident linked to cloud malware[6]. Most of the MACT incidents observed take place in hijacked Microsoft 365 environments[6].
On the detection side, MITRE ATT&CK describes monitoring the use of OAuth tokens by third-party applications to access mail, calendar or SharePoint, when the token was granted recently or following a targeted phishing link[3]. Microsoft recommends enabling "Review admin consent requests" to force the review of new applications[7]. After the January 2024 attack, it also recommends removing applications that hold EWS.AccessAsUser.All or EWS.full_access_as_app if they are no longer needed[8].
Case filesReal cases
September 2022: OAuth applications created on compromised tenants
On September 22, 2022, Microsoft published the analysis of an attack in which malicious OAuth applications were deployed on compromised cloud tenants, to control Exchange Online settings and spread spam[1]. The actor carried out credential stuffing attacks against administrator accounts without multifactor authentication[1]. They registered an application, gave it the Global Administrator and Exchange Online Administrator roles, then added their own credentials to it[1]. The actor grants administrator consent to the application themselves[1].
December 2022: verified publishers imitating legitimate publishers
Proofpoint discovered this campaign on December 6, 2022[5]. It seemed to target mostly organizations and users in the United Kingdom, including finance and marketing staff, managers and executives[5]. Proofpoint informed Microsoft on December 20, 2022, and the campaign ended on December 27, 2022[5]. Microsoft then disabled the applications: those that had been authorized can only access data until the last access token expires, generally between 60 and 90 minutes[5].
July to November 2023: about 17,000 OAuth applications
In an article of December 12, 2023, Microsoft describes an adversary-in-the-middle (AiTM) phishing campaign[7]. The actor created about 17,000 multitenant OAuth applications using compromised accounts[7]. Applications named "oauth" could read emails, and those named "App" could send emails in large volumes[7]. They sent more than 927,000 phishing emails[7]. Microsoft removed all the applications linked to this campaign, which ran from July to November 2023[7].
January 2024: Midnight Blizzard against Microsoft
Microsoft detected an attack on its corporate systems on January 12, 2024 and attributed it to Midnight Blizzard, a Russian state-sponsored actor also known as NOBELIUM[8]. A password spray attack compromised an account in a legacy test tenant without multifactor authentication[8]. The actor then compromised a legacy test OAuth application that had elevated access, and created additional malicious OAuth applications[8]. They created a new user account to grant consent to these applications, then obtained the Exchange Online full_access_as_app role, which gives access to mailboxes[8].
Sources
- Microsoft, “Malicious OAuth applications abuse cloud email services to spread spam”, 22 September 2022. https://www.microsoft.com/en-us/security/blog/2022/09/22/malicious-oauth-applications-used-to-compromise-email-servers-and-spread-spam/ (accessed 14 November 2026)
- Microsoft, “Investigate and remediate risky OAuth apps”. https://learn.microsoft.com/en-us/defender-cloud-apps/investigate-risky-oauth (accessed 14 November 2026)
- MITRE ATT&CK, “Steal Application Access Token”. https://attack.mitre.org/techniques/T1528/ (accessed 14 November 2026)
- Microsoft, “Configure how users consent to applications”. https://learn.microsoft.com/en-us/entra/identity/enterprise-apps/configure-user-consent (accessed 14 November 2026)
- Proofpoint, “The Dangerous Consequences of Threat Actors Abusing Microsoft’s “Verified Publisher” Status”, 31 January 2023. https://www.proofpoint.com/us/blog/cloud-security/dangerous-consequences-threat-actors-abusing-microsofts-verified-publisher (accessed 14 November 2026)
- Proofpoint, “Revisiting MACT: Malicious Applications in Credible Cloud Tenants”, 11 April 2024. https://www.proofpoint.com/us/blog/cloud-security/revisiting-mact-malicious-applications-credible-cloud-tenants (accessed 14 November 2026)
- Microsoft, “Threat actors misuse OAuth applications to automate financially driven attacks”, 12 December 2023. https://www.microsoft.com/en-us/security/blog/2023/12/12/threat-actors-misuse-oauth-applications-to-automate-financially-driven-attacks/ (accessed 14 November 2026)
- Microsoft, “Midnight Blizzard: Guidance for responders on nation-state attack”, 25 January 2024. https://www.microsoft.com/en-us/security/blog/2024/01/25/midnight-blizzard-guidance-for-responders-on-nation-state-attack/ (accessed 14 November 2026)
Cite this article
Brain, “Consent phishing: the app that asks for your permissions”, Cyberpedia, 2026. https:/
BibTeX
@misc{brain_cp012_en_2026,
author = {Brain},
title = {Consent phishing: the app that asks for your permissions},
year = {2026},
howpublished = {Cyberpedia},
url = {https://startbrain.io/cyberpedia/en/articles/oauth-consent-phishing-malicious-apps/},
note = {CC BY 4.0}
}