Incident sheet · /security

Filed
Read
4 min · 709 words
Severity
Routine

Phishing Kits Now Steal Microsoft 365 Session Tokens, Huntress Warns

Huntress researchers analyzed three phishing kits built to capture and replay Microsoft 365 session tokens, bypassing MFA and handing attackers persistent access to cloud tenants.

Inside Modern Phishing Kits Built to Steal Microsoft 365 Sessions
FeedInside Modern Phishing Kits Built to Steal Microsoft 365 Sessions — Shlomaster / pixabay

Event log

  1. Huntress analyzed three Microsoft 365 token-theft kits: EvilTokens, Knight Office and Kali365.
  2. EvilTokens, discovered in early 2026, abused Railway PaaS as a token-harvesting and replay engine.
  3. A joint Microsoft Digital Crimes Unit and FBI operation led to the arrest of an alleged EvilTokens operator.
  4. Kali365 bundles Phishing as a Service into one panel with device code flows, reverse-proxy lures and built-in AI features.

Attackers can now buy phishing kits that steal authenticated Microsoft 365 sessions, complete with admin dashboards and round-the-clock support, researchers at Huntress reported. The kits bypass multi-factor authentication (MFA, an additional login check beyond a password) by capturing session tokens rather than passwords. Even entry-level operators can replay a stolen token into a live session where login and MFA are already complete.

Huntress, a managed security provider, analyzed three kits in detail: EvilTokens, Knight Office and Kali365. All three target Microsoft 365 session tokens — the credentials that keep a user signed in after authentication.

What is inside a modern phishing kit?

A phishing kit is a pre-packaged set of code and templates, with instructions for launching a campaign and exploiting the stolen data. Basic kits ship as simple HTML and script files that render one fake login page.

Advanced kits operate like software-as-a-service businesses. They include admin dashboards that let attackers track victims, modify phishing page content and relay credentials in real time, the Huntress analysis found.

How do attackers turn lures into stolen sessions?

The kits rely on three tradecraft techniques to convert a phishing lure into a usable authenticated session.

  • Device code phishing. The victim authenticates directly with the identity provider while the attacker quietly captures the resulting authorization and reuses it. The victim never types credentials into an unfamiliar page, so the experience feels like a normal sign-in.
  • Adversary-in-the-middle (AiTM) attacks. The attacker relays the authentication flow in real time, either through a fake login page or by intercepting an exchange inside the legitimate app; the victim completes the MFA prompt while the attacker captures the session information.
  • Token replay. After obtaining a sign-in token, a kit such as Kali365 replays it into an active session, giving the attacker access to cloud resources and the victim's email.

These techniques expose the limits of MFA, according to Huntress: a successfully completed MFA prompt does not make a session trustworthy on its own. Token theft also yields more than password theft, because it grants persistent access rather than a single credential.

What did Huntress find in the three kits?

EvilTokens surfaced in early 2026 as part of a device code phishing campaign. It abused Railway, a platform-as-a-service (PaaS, cloud infrastructure for deploying applications) provider, as a token-harvesting and replay engine. A joint operation between Microsoft's Digital Crimes Unit and the FBI later in the year led to the arrest of one of its alleged operators.

Knight Office is an AiTM kit that steals valid Microsoft 365 session tokens and feeds them into an operator console. Attackers can replay those sessions and establish persistence across tenant accounts, Huntress found.

Kali365 packages a full Phishing as a Service (PhaaS, phishing attacks rented as a subscription) operation into a single control panel. Operators can capture tokens through device code flows and reverse-proxy lures, replay them into live Microsoft 365 sessions, maintain access and scale activity with built-in AI features.

The common outcome is persistence: attackers walk away with durable access to tenant accounts, not just usernames and passwords.

What should defenders do about token theft?

A single signal covers only one piece of the attack, Huntress cautioned. Security teams should trace the full sequence from lure to authentication, token issuance or theft, device registration and mailbox activity to understand identity compromise.

Huntress recommends hunting 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.

Response priorities focus on identity artifacts, not just the original lure:

  • Revoke refresh tokens, then remove unauthorized devices and Windows Hello credentials.
  • Restrict or block device code authentication where possible.
  • Investigate mailbox access, forwarding rules and outbound phishing.

Containment is not the end of the process, the researchers stressed. Teams should monitor for token replay or re-entry and continue hunting even after the original phishing domain disappears.

The analysis appears in Huntress's 2026 Cyber Threat Report: State of Phishing, which covers lure themes, impersonated brands and the psychological techniques attackers use to slip past defenses. The report was sponsored and written by Huntress Labs.

via bleepingcomputer.com (Source)

Filed under

  • microsoft-365
  • phishing
  • session-token-theft
  • multi-factor-authentication
  • identity-security
Share this article:

More from Rebecca Stone

Rebecca Stone

Show full bio

Senior reporter covering money, funding and market shifts around enterprise information technology and the leaders who run at CIO Monitor.

199 articles

The newsroom digest

One email a day with the stories that matter. No noise.

By subscribing you agree to the privacy policy.

Cross-references

‹ Previous articleNext article ›