Huntress Details Phishing Kits Stealing Microsoft 365 Tokens
┌─ microsoft-365-token-phishing-kits.md :: hero
Huntress Labs researchers have detailed how three modern phishing kits, EvilTokens, Knight Office, and Kali365, steal Microsoft 365 session tokens, according to a November 17, 2026 report. Rather than collecting only passwords, these kits use device code phishing and adversary-in-the-middle (AiTM) techniques to capture authenticated sessions after multi-factor authentication (MFA) is completed. Attackers who replay stolen tokens gain immediate, persistent access to victims' cloud resources and email accounts, meaning completed MFA does not make a session trustworthy on its own.
Key Facts
| Attribute | Value |
|---|---|
| Affected product | Microsoft 365 |
| Exploited in the wild | Yes (EvilTokens observed in an early 2026 device code phishing campaign; Huntress also analysed Knight Office and Kali365) |
| Disclosure date | November 17, 2026 |
How Do Modern Phishing Kits Steal Microsoft 365 Sessions?
Modern phishing kits capture authenticated Microsoft 365 sessions through device code phishing, adversary-in-the-middle (AiTM) interception, and token replay. Victims sign in and complete multi-factor authentication normally, while attackers quietly capture the resulting authorisation or session tokens and replay them into live sessions to access cloud resources and email.
Basic phishing kits typically rely on simple HTML and script files to host a single fake login page, whereas advanced kits operate like software-as-a-service platforms with admin dashboards and round-the-clock support. Through these consoles, operators track victims, modify lure content, and relay credentials in real time.
In a device code phishing attack, the victim authenticates directly with the identity provider rather than typing credentials into an unfamiliar page, while the attacker captures the resulting authorisation. In an AiTM attack, the attacker relays the authentication flow in real time using a fake login page or by intercepting an exchange within the legitimate application, capturing session data once the victim approves the MFA prompt.
Because stolen sign-in tokens already carry completed login and MFA state, replaying them lets even entry-level attackers enter active Microsoft 365 sessions without needing the victim's password again.
Which Phishing Kits Did Huntress Analyse?
Huntress researchers analysed three phishing kits targeting Microsoft 365 tokens: EvilTokens, Knight Office, and Kali365. Discovered in early 2026, EvilTokens abused the Railway Platform-as-a-Service platform for device code token harvesting, while Knight Office operates as an AiTM console and Kali365 bundles full Phishing as a Service features.
- EvilTokens: Identified by Huntress in early 2026 during a device code phishing campaign that abused Railway Platform-as-a-Service (PaaS) as a token-harvesting and replay engine. A joint effort by Microsoft's Digital Crimes Unit and the FBI later in 2026 led to the arrest of one of its alleged operators.
- Knight Office: An AiTM phishing kit that captures valid Microsoft 365 session tokens and feeds them into an operator console, enabling attackers to replay sessions and establish persistence across tenant accounts.
- Kali365: A Phishing as a Service (PhaaS) kit combining device code flows and reverse-proxy lures in one control panel, allowing operators to capture tokens, replay them into live Microsoft 365 sessions, maintain access, and scale activity using built-in AI features.
A joint operation by Microsoft's Digital Crimes Unit and the FBI later in 2026 led to the arrest of one alleged EvilTokens operator after Huntress uncovered the kit's abuse of the Railway PaaS platform.
How Can Defenders Detect Stolen Session Activity?
Defenders can detect stolen Microsoft 365 sessions by correlating signals across the entire attack sequence, from the initial lure through authentication, token issuance, device registration, and mailbox activity. Investigating only the original phishing page misses how captured tokens are replayed to maintain persistent access inside the tenant.
Huntress advises security teams to start by checking for successful sign-ins originating from unfamiliar cloud infrastructure. Defenders should then hunt for device code authentication events and automated token refreshes, and review newly registered devices, added authentication methods, and unusual mailbox activity.
Tracking the full chain from authentication and token issuance to new device registrations and mailbox changes is necessary to uncover identity compromises caused by session replay.
What Should You Do?
No software patch is involved because these kits exploit legitimate authentication flows and user interaction; organisations should apply the identity hardening and containment measures outlined in the report:
- Restrict device code authentication: Block or restrict device code authentication flows in Microsoft 365 wherever they are not required.
- Monitor sign-ins and token activity: Check for successful logins from unfamiliar cloud infrastructure, automated token refreshes, and newly added devices or authentication methods.
- Contain compromised identities thoroughly: Revoke refresh tokens, then remove any unauthorised devices and Windows Hello credentials rather than addressing only the phishing lure.
- Audit mailbox activity: Investigate suspicious mailbox access, forwarding rules, and outbound phishing sent from affected accounts.
- Continue post-containment hunting: Keep monitoring for token replay or re-entry even after the original phishing domain goes offline.
- Report incidents in Malaysia: Malaysian organisations and individuals who suspect account compromise or phishing can report incidents to Cyber999 (MyCERT) or call the National Scam Response Centre at 997 for scams.
Common Questions
Why does MFA not stop AiTM and device code phishing kits?
In both device code and adversary-in-the-middle attacks, the victim completes the login and MFA challenge directly or in real time through a relay. Because the attacker captures the resulting session token afterwards, the replayed session already carries a verified MFA state.
Why is token theft more valuable to attackers than password theft?
Stolen passwords alone still require attackers to clear login and multi-factor authentication checks. By contrast, replaying a captured Microsoft 365 sign-in token places the attacker directly into an active, already authenticated session with immediate access to cloud resources and email accounts.
What should administrators revoke first after suspected session theft?
Administrators should look beyond blocking the initial phishing lure and contain identity artefacts directly. Key steps include revoking refresh tokens, removing unauthorised devices and Windows Hello credentials, and auditing mailbox access, forwarding rules, and outbound phishing for signs of continued replay.
Sources and Methodology
This dispatch is based exclusively on the November 17, 2026 BleepingComputer article sponsored and written by Huntress Labs, Inside Modern Phishing Kits Built to Steal Microsoft 365 Sessions. No other documents, vendor advisories, or external reports were reviewed or verified for this summary.
$ ls -t news/ | head -5
$ ls ./tags
- #Scams & Phishing