InformationalEcosystem changeMonitoring

Huntress reports phishing kits built to steal Microsoft 365 tokens

Huntress Labs says three phishing kits steal Microsoft 365 session tokens; revoke refresh tokens and restrict device code sign-in if compromised.

Published
Exploitation
Not independently confirmed. Huntress Labs says it found EvilTokens in a device code phishing campaign and analyzed Knight Office and Kali365; no victims or counts are given.

What happened

According to Huntress Labs, which sponsored and wrote the article, modern phishing kits are built to steal Microsoft 365 session tokens rather than just passwords. The piece was published on BleepingComputer as sponsored content on 2026-11-17 (BleepingComputer). This is vendor research, not an independent finding. BleepingComputer is the only outlet carrying it, and no official CERT, government or Microsoft advisory is part of the sources.

Huntress says it analyzed three kits:

  • EvilTokens: Huntress says it found this in early 2026, as part of a device code phishing campaign that abused the Railway PaaS platform as a token-harvesting and replay engine. The article says a joint effort by Microsoft’s Digital Crimes Unit and the FBI later led to the arrest of one of its alleged operators. That arrest is stated only in the Huntress article.
  • Knight Office: Huntress describes it as an adversary-in-the-middle (AiTM) kit that steals valid Microsoft 365 session tokens and feeds them into an operator console. Huntress says attackers can replay those sessions and set up persistence across tenant accounts.
  • Kali365: Huntress describes it as a Phishing as a Service kit with a single control panel. Operators can capture tokens through device code flows and reverse-proxy lures, replay them into live Microsoft 365 sessions, and maintain access. It also has built-in AI features.

Huntress says a replayed token can land an attacker in an active session where login and MFA are already complete. It therefore argues that completed MFA does not make a session trustworthy on its own. The article gives no victim names, no counts and no data on how widely the kits are used. Those details are not confirmed.

Am I affected?

The article does not say any specific organization was compromised. It describes techniques that apply to Microsoft 365 tenants generally. No CVE or software flaw is cited, and no patch applies.

Huntress suggests looking for these signs of a stolen session:

  • Successful sign-ins from unfamiliar cloud infrastructure.
  • Device code authentication and automated token refresh.
  • New devices, new authentication methods, and unusual mailbox activity.

Whether any of these apply to your tenant is not something the sources can tell you.

What to do now

The steps below are the response priorities Huntress lists in its article.

  1. Check for successful sign-ins from unfamiliar cloud infrastructure, then hunt for device code authentication and automated token refresh.
  2. Review new devices, authentication methods and mailbox activity. Follow the full sequence from lure to sign-in, token issuance or theft, device registration and mailbox activity.
  3. If you suspect compromise, revoke refresh tokens first. Then remove unauthorized devices and Windows Hello credentials, since containment must cover identity artifacts and not only the original lure.
  4. Restrict or block device code authentication where you can.
  5. Investigate mailbox access, forwarding rules and outbound phishing.
  6. Keep monitoring for replay or re-entry after containment, and keep hunting even after the original phishing domain disappears.

Sources