Huntress Details Kali365 Phishing Kit’s Microsoft 365 Token Theft and Fraud Tools

· Original article ↗

Summary

Huntress analyzed Kali365/Octopi365, a phishing-as-a-service kit that steals Microsoft 365 tokens through device-code and AiTM attacks, then enables mailbox access, persistence, and follow-on fraud.

Key points

  • Huntress traced a suspected campaign to device-code sign-ins from Tencent Cloud, observing more than 80 successful login events on May 20, 2026.
  • The kit uses Microsoft sign-in lures to capture tokens; in one observed case, access was gained in as little as 42 seconds and could persist after MFA use or password changes.
  • Researchers found more than 240 IPs hosting panel variants, with at least 33 lure templates and over 100 API endpoints across the platform.
  • One edition supports mailbox monitoring and AI-assisted business email compromise; another scans captured mailboxes for sensitive data and can abuse Exchange administration features.
  • Companion desktop tools turn stolen tokens into browser sessions and send phishing emails from compromised accounts, using paced sending and draft management.
  • Huntress recommends investigating device-code activity alongside token use, mailbox access, sign-in logs, user agents, and suspicious infrastructure; it supports restricting device-code authentication where possible.

Article Details

Attack Vectors
  • Phishing emails direct recipients to a lure page and then to a legitimate Microsoft sign-in, where victims are prompted to enter a device code.
  • The E1 panel captures Microsoft 365 sessions through adversary-in-the-middle reverse-proxy cookie theft and device-code-flow lures.
  • Stolen access and refresh tokens are used to open victim mailboxes and establish browser sessions without an interactive sign-in.
  • OctoLink Sender uses compromised accounts to send phishing emails through Microsoft Graph, including to other recipients in the victim's organization.
  • The E2 edition supports follow-on business email compromise by analyzing intercepted conversations and drafting replies for operators to send from compromised mailboxes.
Defensive Notes
  • Microsoft recommends that organizations block device-code flow as broadly as possible, allowing exceptions only where there is a legitimate need.
  • Investigate device-code sign-ins as possible evidence of broader token abuse, mailbox access, persistence, and follow-on activity.
  • Hunt across sign-in logs, token activity, mailbox use, unusual user agents, and suspicious infrastructure. The article notes that OctoLink apps can create real browser sessions using SSO cookies rather than scripted Graph traffic.
  • The article's detection guidance includes reviewing sign-ins from AS132203 or 43.173.64.0/20 and the user agent, while noting that residential proxies can complicate IP-based detection.

MITRE ATT&CK

People

Vendors

Products

Tools

Related Articles