# Google Session Hijacking Campaign via Calendar Invite Lures and Live Reverse Proxy (Operation CloudFire, Nyasher)

> Material Security tracks 'Operation CloudFire', a phishing campaign active since mid-August 2026 that sends lures as Google Calendar invites, routes victims through Google-hosted redirectors, and uses a live reverse proxy (the Nyasher framework) in front of the real Google sign-in page to relay passwords and 2FA in real time and capture authenticated sessions. Hijacked sessions give access to mail, Drive, and the ability to grant OAuth access in the victim's name.

- **Published:** 2026-11-14T11:32:20Z
- **Last reviewed:** 2026-11-14T11:32:20Z
- **Canonical:** https://intel.threadlinqs.com/threat/TL-2026-6072
- **Card image:** https://intel.threadlinqs.com/og/threat/TL-2026-6072.png
- **ID:** TL-2026-6072
- **Severity:** HIGH
- **Category:** PHISHING
- **Status:** MONITORING
- **Detections:** 9 (1 shown free below) · **IOCs:** 8 (2 shown free below)

## Description

Operation CloudFire is a credential-and-session theft campaign reported by Material Security's Threat Research team (vendor-reported; BleepingComputer coverage is sponsored/written by Material Security). The campaign has been observed since mid-August 2026 and, per The Hacker News (2026-11-12), appears to have wound down for now, with the researchers expecting a rebrand. Lures themed as voicemail notifications, RFP/RFI requests, court summonses and missed-payment notices arrive as Google Calendar invitations (sometimes BCC'd), frequently from organizations the recipient already corresponds with because the senders are compromised legitimate accounts that mail their owners' saved contacts. Send timing tracked US business hours.

Links in the lures mostly pass through Google's own /url redirector with the real destination in a query parameter, giving the first hop a trusted Google URL. Other waves used share.google links and Google Drive file links, and one variant used Intuit transaction-notification links. The victim then lands on an attacker gateway fronted by a Cloudflare Turnstile challenge (in some builds the genuine widget served from Cloudflare infrastructure). After the challenge passes, the password-capture page is built client-side by JavaScript, so no static page exists for crawlers to scan.

Behind the gateway a live reverse proxy sits in front of the genuine Google sign-in. The victim types credentials and approves prompts while the operator watches in real time. The proxy relays SMS codes, authenticator-app codes, Google's Tap-Yes number matching, and phone-based account recovery; none of these stop session capture because the authenticated session is taken after the victim completes MFA. The framework is Nyasher, publicly documented since February 2026 after researchers found a hardcoded key in its JavaScript. An unfinished Microsoft 365 branch was found in the code (stylesheet path, second harness, and a title string that decodes to a Microsoft sign-in prompt), though none of those paths currently resolve.

Infrastructure is disposable: many lure domains across several registrars, mostly registered days before use; receiving endpoints went live 1-20 days before the wave they served; and one wildcard certificate was issued 46 minutes before its domain was registered. Material reported gateway and backend URLs to Cloudflare's abuse system; Cloudflare confirmed them and placed phishing interstitials in front of them, forcing the operator to re-register, re-point and re-certify infrastructure. In one month of data from Material customers that share data (a lower bound), Material reports 68,676 campaign messages blocked, 2,112 unique malicious URLs, 897 domains taken down, and 8,570 compromised sender addresses across 6,354 root domains. Lure themes have since shifted to Google Drive file/document share invites.

Post-compromise, a live session exposes mail and Drive and allows the operator to grant OAuth access in the victim's name, which can persist after a password reset unless the grant is revoked. Hijacked sessions appear in Google Workspace audit logs. No threat actor has been named, and no specific domains, IPs or hashes are published in the public articles; Material points to a companion whitepaper (September 2026) for IOCs, which could not be read as text during this research. Severity is analyst judgment.

## MITRE ATT&CK

- T1566.002 Spearphishing Link
- T1204.001 Malicious Link
- T1557 Adversary-in-the-Middle
- T1111 Multi-Factor Authentication Interception
- T1539 Steal Web Session Cookie
- T1550.004 Web Session Cookie
- T1078.004 Cloud Accounts
- T1583.001 Domains
- T1586.002 Email Accounts
- T1090 Proxy
- T1671 Cloud Application Integration
- T1114 Email Collection
- T1530 Data from Cloud Storage

## Free detection (Splunk SPL, Microsoft KQL, Sigma)

**Operation CloudFire: Google Workspace Login Followed by OAuth Grant from a Second IP (Splunk)** (`TL-2026-6072-SPL-002`)

After the Nyasher live proxy relays the victim's password and MFA, the operator replays the captured session from proxy-origin infrastructure and can grant OAuth access that survives a password reset. This search correlates Workspace login_success and token authorize events per user within 30 minutes and alerts when more than one source IP is involved (victim IP vs operator IP). Maps to T1539, T1550.004, T1078.004 and T1671.

- **Languages:** Splunk SPL, Microsoft KQL, Sigma
- **ATT&CK:** T1539, T1550.004, T1078.004, T1671
- **Data sources:** gws:reports:login, gws:reports:token, CloudAppEvents, google_workspace
- **Severity:** high
- **Confidence:** medium

### Splunk SPL

```spl
index=gws (sourcetype="gws:reports:login" OR sourcetype="gws:reports:token") earliest=-4h
| eval user=lower('actor.email'), evname=mvindex('events{}.name',0)
| eval kind=case(sourcetype=="gws:reports:login" AND evname=="login_success","login", sourcetype=="gws:reports:token" AND evname=="authorize","oauth_grant")
| where isnotnull(kind)
| stats values(eval(if(kind=="login",ipAddress,null()))) AS login_ips values(eval(if(kind=="oauth_grant",ipAddress,null()))) AS grant_ips min(eval(if(kind=="login",_time,null()))) AS t_login max(eval(if(kind=="oauth_grant",_time,null()))) AS t_grant BY user
| where isnotnull(t_login) AND isnotnull(t_grant) AND t_grant>=t_login AND (t_grant-t_login)<=1800
| eval all_ips=mvdedup(mvappend(login_ips,grant_ips))
| where mvcount(all_ips)>1
| convert ctime(t_login) ctime(t_grant)
```

### Microsoft KQL

```kql
CloudAppEvents
| where Timestamp > ago(4h)
| where Application has 'Google Workspace'
| extend Kind = case(ActionType has_any ('login_success', 'Log on', 'Logon'), 'login', ActionType has_any ('authorize', 'OAuth'), 'oauth_grant', '')
| where Kind != ''
| summarize LoginIPs = make_set_if(IPAddress, Kind == 'login'), GrantIPs = make_set_if(IPAddress, Kind == 'oauth_grant'), FirstLogin = minif(Timestamp, Kind == 'login'), LastGrant = maxif(Timestamp, Kind == 'oauth_grant') by AccountId, Window = bin(Timestamp, 30m)
| where array_length(LoginIPs) > 0 and array_length(GrantIPs) > 0
| extend AllIPs = set_union(LoginIPs, GrantIPs)
| where array_length(AllIPs) > 1
| project Window, AccountId, LoginIPs, GrantIPs, FirstLogin, LastGrant
```

### Sigma

```yaml
title: Google Workspace OAuth Grant With Mail or Drive Scope (Operation CloudFire Persistence)
id: 0b9d7e44-81c2-4f6a-b3a5-2e1f9c6d4a77
status: experimental
description: Detects Google Workspace token authorize events granting Gmail or Drive scopes to a third-party app. Operation CloudFire operators use a hijacked session to grant OAuth access in the victim's name, which persists after a password reset. Correlate with a preceding login from a different IP.
references:
  - https://thehackernews.com/2026/11/calendar-invites-and-live-proxies.html
author: Threadlinqs Team
date: 2026-11-14
tags:
  - attack.persistence
  - attack.t1671
  - attack.defense-evasion
  - attack.t1550.004
  - attack.initial-access
  - attack.t1078.004
logsource:
  product: google_workspace
  service: google_workspace.token
detection:
  selection:
    eventName: authorize
  selection_scope:
    scope|contains:
      - 'mail.google.com'
      - 'auth/gmail'
      - 'auth/drive'
  condition: selection and selection_scope
falsepositives:
  - Users authorizing approved productivity or backup applications
level: medium
```

**Known false positives:**

- Users who travel or switch between mobile and VPN networks and then authorize a legitimate app
- Admin-approved OAuth apps being authorized by many users during rollout

**Tuning notes:** Allowlist approved OAuth client IDs and corporate egress/VPN ranges. Field names depend on your Workspace ingestion (Splunk Add-on for Google Workspace flattening; Defender for Cloud Apps ActionType values) so verify them against sample events. Map the Sigma logsource to the Reports API 'token' application in your pipeline; this rule alone has no IP correlation, so pair it with the SPL/KQL form. Extend the window or add a new-device check for higher fidelity.

### 8 more detection rules (Splunk SPL, Microsoft KQL, Sigma), unlock with TL Creds

- Operation CloudFire: Calendar-Invite Email with [indicator]/url Redirect to Non-Google Host (KQL) (locked: unlock with TL Creds)
- Operation CloudFire: Google Workspace Session Replay and OAuth Grant (KQL) (locked: unlock with TL Creds)
- Operation CloudFire: Endpoint Reaches Turnstile and Google Lookalike Host (KQL) (locked: unlock with TL Creds)
- Operation CloudFire: Google Redirector Abuse to Non-Google Destination (Sigma) (locked: unlock with TL Creds)
- Operation CloudFire: Google Workspace OAuth Grant With Mail or Drive Scope (Sigma) (locked: unlock with TL Creds)
- Operation CloudFire: DNS Query to Google Sign-in Lookalike Domain (Sigma) (locked: unlock with TL Creds)
- Operation CloudFire: [indicator]/url Redirector to Non-Google Destination (Splunk) (locked: unlock with TL Creds)
- Operation CloudFire: Turnstile-Gated Google Sign-in Lookalike Domain Access (Splunk) (locked: unlock with TL Creds)

Sign up free to unlock the detection rules: https://threadlinqs.com/?modal=signup

## Sample indicators (2 of 8)

Defanged, TLP:CLEAR.

- Domain: `google[.]com`
- Domain: `share[.]google`

Full set of 8 indicators unlocks with TL Creds: https://threadlinqs.com/tl-creds

## Sources

- [Calendar Invites and Live Proxies: Inside a Google Session Hijacking Campaign](https://thehackernews.com/2026/11/calendar-invites-and-live-proxies.html)
- [Google Calendar phishing campaign hijacks sessions, bypasses MFA (BleepingComputer, sponsored by Material Security)](https://bleepingcomputer.com/news/security/google-calendar-phishing-campaign-hijacks-sessions-bypasses-mfa)
- [Material Security companion whitepaper: Operation CloudFire calendar invite campaign (IOCs and timeline)](https://content.material.security/hubfs/Material%20Security%20Site%20CMS/CalendarInviteWhitepaper-sep2026.pdf)
- [The Quiet Phish: Stopping Calendar Invitation Attacks (Material Security)](https://material.security/resources/the-quiet-phish-stopping-calendar-invitation-attacks)
- [Operation CloudFire phishing campaign targets corporate Google sessions (SC Media brief)](https://www.scworld.com/brief/operation-cloudfire-phishing-campaign-targets-corporate-google-sessions)
- [Google Calendar phishing campaign hijacks sessions and bypasses MFA (Paubox)](https://www.paubox.com/blog/google-calendar-phishing-campaign-hijacks-sessions-and-bypasses-mfa)

## Full data

The other 8 detection rules (Splunk SPL / Microsoft KQL / Sigma) unlock with TL Creds; a free account includes daily creds: https://threadlinqs.com/?modal=signup
Indicator values unlock with TL Creds, at a lower price on Red and Purple; AI agents can query rules and IOCs through the Threadlinqs MCP server (Purple): https://intel.threadlinqs.com/mcp

Canonical: https://intel.threadlinqs.com/threat/TL-2026-6072
