Skip to navigation

Threat reportPhishingTL-2026-5957

Phishing Surge in Europe: 4.5x Growth vs 1.5x Globally, Targeted Cloud Account Takeover and Callback Phishing (Proofpoint)

mediumACTIVE

Proofpoint telemetry (July 2025 - September 2026) shows phishing detections up 4.5x across Europe versus 1.5x globally, with a six-fold rise in the UK. European users clicked 8.3 of every 100 delivered threat messages (2.0 globally), and 64% of European organizations hit by precision cloud attacks showed evidence of successful compromise.

Key facts

IdentifiersTL-2026-5957
Also known asEurope's Threat Landscape 2026
SeverityMEDIUM
First disclosed
Affected productsVarious Cloud email and identity platforms (Microsoft 365 / webmail incl. Outlook, Zimbra); Open source Langflow and LiteLLM (AI infrastructure targeted by exploitation attempts)
AttributionTA488, TA2725, UAC-0050, UNK_AcademicFlare (medium confidence)
MalwareAstaroth, Evilginx, FluxEngine RAT, Grandoreiro - S0531, Metamorfo, Mispadu - S1122
ToolsEvilginx, Remote Manipulator System
MITRE ATT&CK techniquesT1078.004 Valid Accounts: Cloud Accounts, T1098 Account Manipulation, T1110 Brute Force, T1110.003 Brute Force: Password Spraying, T1111 Multi-Factor Authentication Interception, T1114.003 Email Collection: Email Forwarding Rule, T1190 Exploit Public-Facing Application, T1219 Remote Access Tools, T1528 Steal Application Access Token, T1539 Steal Web Session Cookie, T1550.001 Use Alternate Authentication Material: Application Access Token, T1557 Adversary-in-the-Middle and 5 more in the attack flow
MotivationUNKNOWN
Target sectorshealth, hospitality, transportation-logistics, finance, electronics, government administration, defense, education, think tanks
Target regionsEurope, united kingdom, 155 - Western Europe, ukraine, spain, germany, france
Detection rules9 (1 shown free)
Sources5 cited sources

Detection coverage

As of 2026-11-13, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-5957 across Splunk SPL, Microsoft KQL and Sigma, covering 21 indicator(s) of compromise. The whole corpus is readable without an account. Every threat comes with one complete detection in Splunk SPL, Microsoft KQL and Sigma, free for everyone; a free account adds daily TL Creds to unlock the rest. Paid plans add a larger daily allowance, correlation and the MCP server. Threadlinqs MCP server · View plans.

Free detection: Entra ID Device-Code Flow Sign-in (Device-Code and ASP Phishing) — Exploitation

TL-2026-5957 (Proofpoint Europe Threat Landscape 2026): Russian-aligned actors use Device-code and App-Specific Password (ASP) phishing kits that reach inboxes without victims entering credentials. This detection alerts on successful Entra ID sign-ins that used the OAuth device-code flow (T1528, T1078.004, T1566.002), which in most tenants is rare for ordinary users and is the end state of a device-code phish. Lists first/last seen, apps, IPs and countries per user for triage.

  • Languages: Splunk SPL, Microsoft KQL, Sigma
  • ATT&CK: T1078.004, T1528, T1566.002
  • Data sources: azure:monitor:aad SignInLogs, SigninLogs
  • Severity: high
  • Confidence: medium

Splunk SPL

index=azure sourcetype="azure:monitor:aad" category=SignInLogs ("properties.authenticationProtocol"=deviceCode OR "properties.originalTransferMethod"=deviceCodeFlow) "properties.status.errorCode"=0
| eval user='properties.userPrincipalName', src_ip='properties.ipAddress', app='properties.appDisplayName', country='properties.location.countryOrRegion'
| stats count AS signins min(_time) AS first_seen max(_time) AS last_seen values(app) AS apps values(src_ip) AS src_ips values(country) AS countries BY user
| convert ctime(first_seen) ctime(last_seen)

Microsoft KQL

SigninLogs
| where TimeGenerated > ago(24h)
| where ResultType == '0'
| where AuthenticationProtocol =~ 'deviceCode' or OriginalTransferMethod =~ 'deviceCodeFlow'
| project TimeGenerated, UserPrincipalName, IPAddress, Location, AppDisplayName, ResourceDisplayName, UserAgent, ConditionalAccessStatus, CorrelationId

Sigma

title: Entra ID Successful Sign-in Using Device Code Flow (Device-Code Phishing)
id: 5b1c7e42-3a9d-4f08-9c61-2d84e0a7b1f3
status: experimental
description: Detects successful Entra ID sign-ins via the OAuth device code flow, the end state of device-code phishing kits used by Russian-aligned actors (Proofpoint Europe Threat Landscape 2026, TL-2026-5957).
references:
  - https://www.proofpoint.com/uk/blog/threat-insight/europes-threat-landscape-2026-espionage-expands-ai-and-phishing-risks-grow
author: Threadlinqs Team
date: 2026/11/13
tags:
  - attack.initial-access
  - attack.t1078.004
  - attack.credential-access
  - attack.t1528
  - attack.t1566.002
logsource:
  product: azure
  service: signinlogs
detection:
  selection:
    AuthenticationProtocol: deviceCode
    ResultType: '0'
  condition: selection
falsepositives:
  - Legitimate use of device code flow by shared devices, TVs, Teams Rooms, CLI tools (Azure CLI, PowerShell) and IoT
level: high

Known false positives:

  • Legitimate device-code sign-ins from Azure CLI, PowerShell, Teams Rooms, smart TVs or kiosks
  • Developers authenticating headless servers or CI tooling

Tuning notes: Baseline legitimate device-code users/apps for 30 days and allowlist them; alert with higher priority when the sign-in IP/country is new for the user or the app is Microsoft Office / Authentication Broker. Pair with a Conditional Access policy blocking device code flow where not required. Adjust the index/sourcetype to your Microsoft Cloud Services add-on configuration.

8 more detection rules (Splunk SPL, Microsoft KQL, Sigma) · unlock with TL Creds. Sign up free to unlock the detection rules or query them with the Threadlinqs MCP server (Purple)
  • AiTM / Evilginx Session Hijack Risk Detections — Exploitation · unlock with TL Creds
  • OAuth Consent Grant to App with High-Risk Mail/File Scopes — Post-Exploitation · unlock with TL Creds
  • Langflow/LiteLLM Host Spawning Shell or Download Utility — Exploitation of AI Infrastructure · unlock with TL Creds
  • Entra ID Password Spraying Against Many Accounts — Exploitation · unlock with TL Creds
  • Authentication Method / Security Info Change on Account — Post-Exploitation · unlock with TL Creds
  • DNS Query to Community-Reported Evilginx-Tagged Phishing Infrastructure — C2/Infrastructure · unlock with TL Creds
  • Suspicious Mailbox Forwarding/Hiding Inbox Rule After Account Takeover — Post-Exploitation · unlock with TL Creds
  • Remote Manipulator System (RMS) and Remote-Assist Tool Execution (UAC-0050 / TOAD) — C2 · unlock with TL Creds

Indicators of compromise

As of 2026-11-13, Threadlinqs Intelligence tracks 21 indicators of compromise for TL-2026-5957.

21 indicators of compromise · unlock with TL Creds. Compare plans

Technical analysis

Proofpoint's 'Europe's Threat Landscape 2026' report (released 2026-11-12, presented by VP of threat research Daniel Blackford at Proofpoint Protect London) is trend-level intelligence rather than a single-exploit threat. Phishing detections in Europe rose 4.5x year over year versus 1.5x worldwide (3.2x per million inbound messages), with a six-fold rise in the UK. Europeans clicked 8.3 of every 100 delivered threat messages versus 2.0 globally.

Cloud account takeover is the main risk vector. 99% of European organisations were targeted by brute-force attacks (33% impacted), while precision attacks using phishing and session hijacking targeted 98% and impacted 64%; 72% of organisations showed post-access activity, such as manipulating authentication controls, creating mailbox rules, linking apps, and accessing files. Proofpoint documents real-time phishing kits where a live operator captures MFA codes, adversary-in-the-middle (AiTM) kits including Evilginx, device-code and App-Specific Password (ASP) phishing that reaches inboxes without victims entering credentials, and abuse of OAuth consent grants, stolen tokens and overprivileged apps.

Sector findings: callback phishing (TOAD) rose 257% in healthcare (malware detections +84%) and 262% in hospitality, where malware made up 21% of threats (11% worldwide). Electronics firms saw BEC/impostor messages at 3.4x the European average, often as supply-chain invoice fraud. Logistics and freight firms were targeted with RATs and TIMOCOM impersonation to enable cargo theft. TA2725 delivered Latin American banking malware (Mispadu, Metamorfo, Astaroth, Grandoreiro, and the newly observed FluxEngine RAT) with Spanish and Portuguese lures and geofenced payloads. TA2730 used real-time phishing kits and W-8BEN tax-form lures impersonating Swissquote, Binance and Interactive Brokers. A credential phishing campaign impersonating the French health insurer Ameli used hundreds of malicious domains and a previously unobserved kit.

State-aligned activity is expanding: South Asian actor TA395 is moving into East Asia and Western Europe, TA488 (Russian) uses half-click exploits against Outlook, Zimbra and other webmail, UNK_AcademicFlare targets European government, education and think tanks, Russian actors use device-code and ASP kits, and Chinese state-aligned actors use browser/webmail exploitation and AiTM kits against government and defence and, increasingly, AI policy targets. Proofpoint assesses LLM use has accelerated development by several of these actors. On AI infrastructure, more than 188,000 AI-enabled apps were seen connected via OAuth/authentication frameworks, and the share of European sensors hit by attacks on open-source AI tools (Langflow, LiteLLM) rose 4.6x between consecutive 90-day periods. The sources provide no network IOCs (IPs, domains, hashes); no BeaconBeagle correlation was possible and none is claimed. No CVEs are named.

MITRE ATT&CK techniques by tactic:

Initial Access

T1078.004 Valid Accounts: Cloud Accounts; T1190 Exploit Public-Facing Application; T1566.002 Phishing: Spearphishing Link; T1566.004 Phishing: Spearphishing Voice

Persistence

T1098 Account Manipulation

Credential Access

T1110 Brute Force; T1110.003 Brute Force: Password Spraying; T1111 Multi-Factor Authentication Interception; T1528 Steal Application Access Token; T1539 Steal Web Session Cookie; T1557 Adversary-in-the-Middle

Collection

T1114.003 Email Collection: Email Forwarding Rule

Command and Control

T1219 Remote Access Tools

lateral-movement

T1550.001 Use Alternate Authentication Material: Application Access Token

Resource Development

T1583.001 Acquire Infrastructure: Domains; T1588.002 Obtain Capabilities: Tool

Defense Evasion

T1684.001 Impersonation

Remediation

Patches

  • Keep webmail platforms (Outlook, Zimbra) and open-source AI tooling (Langflow, LiteLLM) on current vendor-supported versions

Immediate actions

  • Review and revoke suspicious OAuth consent grants, stolen/overprivileged app tokens and unrecognised connected AI-enabled apps
  • Hunt for new inbox/forwarding rules, new app links and MFA/authentication-method changes following suspicious sign-ins
  • Block or restrict device-code authentication flow and App-Specific Password creation where not required
  • Alert healthcare and hospitality staff to callback (TOAD) phishing and route phone-based requests through verified numbers

Workarounds

  • Restrict internet exposure of open-source AI tools such as Langflow and LiteLLM
  • Require admin consent for third-party app OAuth grants

Longer-term hardening

  • Deploy phishing-resistant MFA (FIDO2/passkeys) and token-binding or conditional access to resist AiTM and session hijacking
  • Govern OAuth app consent with admin approval workflows and least-privilege scopes
  • Harden webmail and browser patching (Outlook, Zimbra) and monitor post-access behaviour in cloud tenants
  • Run localised (language-specific) phishing awareness training

Affected and fixed versions

  • Various — Cloud email and identity platforms (Microsoft 365 / webmail incl. Outlook, Zimbra)
    Vulnerable versions: Tenants without phishing-resistant MFA or OAuth consent governance
  • Open source — Langflow and LiteLLM (AI infrastructure targeted by exploitation attempts)
    Vulnerable versions: Internet-exposed deployments

Timeline

  • UAC-0050 (DaVinci Group) active since at least 2020 (year-granular in the source), using RMM tools such as RMS against Ukrainian government, financial and public-sector targets.
  • From mid-2025 onward TA2725 appears in nearly every observed campaign targeting Spain, delivering Latin American banking malware with geofenced delivery (month-granular in the source).
  • Start of the Proofpoint telemetry window (July 2025) used to measure European phishing growth; day is month-granular in the source.
  • French authorities disclosed a taxpayer data breach affecting hundreds of thousands of citizens; ANSSI attributed it to poor security practices at the tax administration. Report cites it as context for French-targeted phishing; year inferred from report context.
  • Proofpoint researchers identified a credential phishing campaign impersonating the French National Health Insurance (Ameli), using hundreds of malicious domains, Cloudflare filtering, a novel CAPTCHA and a previously unseen kit; year inferred from report context.
  • End of the Proofpoint telemetry window (September 2026); day is month-granular in the source.
  • Daniel Blackford (VP threat research) presents findings at Proofpoint Protect London, including the 64% successful-compromise rate for precision cloud attacks.
  • Proofpoint publishes 'Europe's Threat Landscape 2026', reporting 4.5x phishing growth in Europe vs 1.5x globally and a six-fold rise in the UK.
  • Infosecurity Magazine reports the findings, including healthcare callback phishing +257% and the 188,000+ OAuth-connected AI-enabled apps.

Sources cited

More from TA488

Previous and next in phishing

Drafted by the Threadlinqs research pipeline from 5 sources; published ; last updated —.

Threadlinqs Intelligence — Real-Time Threat Detection Platform

// threat_feed
$ sort --newest
Showing all threats

Live intelligence console

Threat level

Latest Threats