Threat reportAPTTL-2026-0131
AI-Augmented FortiGate Mass Exploitation — Russian-Speaking Actor Breaches 600+ Firewalls Across 55 Countries Using LLM-Generated Tooling and Custom MCP Framework
Amazon Threat Intelligence disclosed a campaign by a Russian-speaking, financially motivated threat actor who breached 600+ FortiGate firewalls across 55+ countries between January 11 and February 18, 2026. The actor used brute-force attacks against exposed management interfaces (no zero-days) then leveraged multiple commercial LLM services (DeepSeek, Claude) to generate reconnaissance tools, automate credential extraction, plan lateral movement, and build a custom MCP server (ARXON) integrating AI directly into the kill chain.
Key facts
| Identifiers | TL-2026-0131, CVE-2019-7192, CVE-2023-27532, CVE-2024-40711 |
|---|---|
| Also known as | AI-Augmented FortiGate Campaign, ARXON Campaign, CHECKER2 Campaign |
| Severity | CRITICAL |
| Exploitation status | Linked CVE exploited in the wild; Listed in CISA KEV: CVE-2019-7192 (added ), CVE-2023-27532 (added ), CVE-2024-40711 (added ) |
| EPSS | CVE-2024-40711: 90.3% (99.7th percentile); CVE-2019-7192: 88.1% (99.7th percentile); CVE-2023-27532: 81.3% (99.6th percentile) |
| Affected products | Fortinet FortiGate; Veeam Backup & Replication; QNAP NAS; Microsoft Active Directory |
| Weaknesses (CWE) | CWE-521, CWE-306, CWE-522, CWE-798 |
| Nation-state nexus | Russia |
| Tools | ARXON MCP Server, CHECKER2, HexStrike, gogo (port scanner) |
| MITRE ATT&CK techniques | T1003.001 LSASS Memory, T1003.006 DCSync, T1016 System Network Configuration Discovery, T1018 Remote System Discovery, T1021.002 SMB/Windows Admin Shares, T1041 Exfiltration Over C2 Channel, T1046 Network Service Discovery, T1059.001 PowerShell, T1059.004 Unix Shell, T1059.006 Python, T1069.002 Domain Groups, T1070 Indicator Removal and 20 more in the attack flow |
| Motivation | FINANCIAL |
| Target sectors | government, telecommunications, manufacturing, media, healthcare, managed-service-providers, enterprise, critical-infrastructure |
| Target regions | South Asia, Latin America, Caribbean, West Africa, Northern Europe, Southeast Asia, Turkey, South Korea, Egypt, Vietnam, Kenya, Asia-Pacific |
| Detection rules | 9 (1 shown free) |
| Sources | 12 cited sources |
Detection coverage
As of 2026-02-22, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-0131 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: FortiGate Mass Brute-Force, Config Extraction, and VPN Credential Recovery
Detects initial access: mass brute-force against FortiGate management on ports 443/8443/10443/4443, configuration backup extraction after successful authentication, and SSL-VPN credential recovery from stolen configs. Campaign hit 600+ devices across 55 countries in 38 days. Staging IPs: 212.11.64.250 (AS4264 Zurich), 185.196.11.225 (Kali).
- Languages: Splunk SPL, Microsoft KQL, Sigma
- ATT&CK: T1110.003, T1078.001, T1602.002, T1552.001, T1133
- Data sources: fortigate_logs, firewall_auth_logs, vpn_logs
- Severity: critical
- Confidence: high
Splunk SPL
index=fortinet sourcetype="fortigate:log"
| eval is_auth_fail=if(action="login" AND status="failed", 1, 0)
| eval is_auth_success=if(action="login" AND status="success", 1, 0)
| eval is_config_export=if(match(action, "(backup|config_download|get_system_config)"), 1, 0)
| eval is_mgmt_port=if(dest_port IN (443, 8443, 10443, 4443), 1, 0)
| where (is_auth_fail=1 OR is_auth_success=1 OR is_config_export=1) AND is_mgmt_port=1
| stats count(eval(is_auth_fail=1)) as fail_count count(eval(is_auth_success=1)) as success_count count(eval(is_config_export=1)) as config_exports by src_ip, dest_ip
| where fail_count > 10 OR config_exports > 0
| eval alert=case(config_exports > 0 AND fail_count > 10, "brute_force_then_config_dump", config_exports > 0, "config_extraction", fail_count > 10, "brute_force")
| table src_ip, dest_ip, fail_count, success_count, config_exports, alertMicrosoft KQL
let auth_events = CommonSecurityLog
| where Timestamp > ago(24h)
| where DeviceVendor == "Fortinet"
| where Activity has_any ("login", "authentication")
| extend IsFailure = Activity has "failed"
| extend IsSuccess = Activity has "success"
| extend IsMgmtPort = DestinationPort in (443, 8443, 10443, 4443);
let config_events = CommonSecurityLog
| where Timestamp > ago(24h)
| where DeviceVendor == "Fortinet"
| where Activity has_any ("backup", "config_download", "get_system_config")
| extend IsConfigExport = true;
let brute = auth_events
| where IsMgmtPort
| summarize FailCount = countif(IsFailure), SuccessCount = countif(IsSuccess) by SourceIP, DestinationIP, bin(Timestamp, 5m)
| where FailCount > 10
| extend AlertType = "fortigate_brute_force";
let config = config_events
| extend AlertType = "config_extraction"
| project Timestamp, SourceIP, DestinationIP, Activity, AlertType;
union brute, config
| project Timestamp, SourceIP, DestinationIP, FailCount, SuccessCount, AlertType
| sort by Timestamp descSigma
title: Mass Brute-Force Against FortiGate Management Interface with Config Extraction
id: 886b8026-f163-4488-a952-fde998c1ad2e
status: experimental
description: Detects mass authentication attempts on FortiGate management ports followed by configuration backup export. AI-augmented campaign targeting 600+ devices across 55 countries.
references:
- https://www.forescout.com/blog/ai-augmented-attacks-against-fortinet-devices/
author: Threadlinqs Team
date: 2026/02/22
tags:
- attack.credential_access
- attack.t1110.003
- attack.collection
- attack.t1602.002
- attack.initial_access
- attack.t1078.001
logsource:
product: fortinet
category: firewall
detection:
selection:
action: 'login'
status: 'failed'
dst_port|any:
- 443
- 8443
- 10443
- 4443
condition: selection | count(src_ip) by dst_ip > 10
timeframe: 5m
falsepositives:
- Legitimate admin with wrong password
- Password rotation causing temporary failures
level: criticalKnown false positives:
- Legitimate admin with wrong password
- Password rotation causing temporary auth failures
Tuning notes: Threshold 10 failures in 5min per source. Correlate successful auth + config export for highest fidelity. Add staging IPs (212.11.64.250, 185.196.11.225) to IOC watchlist.
- FortiGate Brute-Force and CHECKER2 Docker-Orchestrated Parallel VPN Scanning · unlock with TL Creds
- Active Directory Compromise Chain and Veeam Pre-Ransomware Targeting · unlock with TL Creds
- ARXON MCP AI Attack Planning, Staging Server Exfiltration, and Offensive Tool Transfer · unlock with TL Creds
- FortiGate Configuration Backup Extraction After Authentication (Credential Theft) · unlock with TL Creds
- Pass-the-Hash Lateral Movement via Impacket psexec/wmiexec (NTLM Without Kerberos TGT) · unlock with TL Creds
- BloodHound AD Collection and Offensive Tool Execution (Impacket/Nuclei/gogo) · unlock with TL Creds
- Domain Compromise via DCSync, NTLM Relay, Pass-the-Hash, and Veeam Credential Extraction · unlock with TL Creds
- ARXON/CHECKER2 AI Infrastructure, Staging Server Communication, and FortiGate VPN Backdoor · unlock with TL Creds
Indicators of compromise
As of 2026-02-22, Threadlinqs Intelligence tracks 21 indicators of compromise for TL-2026-0131.
Sample indicators (2 of 21)
- IP address:
212.11.64[.]250 - IP address:
185.196.11[.]225
Defanged, TLP:CLEAR. Full set of 21 indicators: unlock with TL Creds.
Technical analysis
Post-access activity targeted SSL-VPN credentials, Active Directory via DCSync, and Veeam backup infrastructure — a classic pre-ransomware playbook. A misconfigured staging server (212.11.64.250) exposed 1,402 files providing unprecedented visibility into AI-assisted attack methodology.
## Overview
On February 21, 2026, Amazon Integrated Security (CJ Moses, CISO) published research documenting a Russian-speaking threat actor who compromised 600+ FortiGate firewall appliances across 55+ countries in just 5 weeks (January 11 — February 18, 2026). The campaign was independently corroborated by Cyber and Ramen (Hunt.io research) and reported by BleepingComputer.
This campaign is significant not for sophisticated exploitation — no zero-days were used — but for demonstrating how commercial AI services enable low-to-medium skill actors to achieve nation-state-scale operations. The actor used at least two commercial LLM providers (DeepSeek and Anthropic Claude) throughout every phase of the kill chain: tool development, attack planning, reconnaissance automation, credential extraction, lateral movement planning, and operational documentation.
## Initial Access: Mass Credential Abuse
The actor's initial access was brute-force attacks against FortiGate management interfaces exposed to the internet. Systematic scanning targeted ports 443, 8443, 10443, and 4443, followed by authentication attempts using commonly reused credentials. No FortiGate vulnerabilities were exploited for initial access — the campaign succeeded entirely through exposed management ports and weak/single-factor credentials.
FortiGate configuration files were high-value targets containing: SSL-VPN user credentials with recoverable passwords, administrative credentials, complete network topology and routing information, firewall policies revealing internal architecture, and IPsec VPN peer configurations. The actor developed AI-assisted Python scripts to parse, decrypt, and organize stolen configurations.
## Geographic Distribution
The campaign's targeting was opportunistic rather than sector-specific, consistent with automated mass scanning. Concentrations of compromised devices were observed across South Asia, Latin America, the Caribbean, West Africa, Northern Europe, and Southeast Asia. Patterns of contiguous IP blocks and shared non-standard management ports indicated MSP deployments and large organizational networks.
Confirmed compromises included an industrial gas company in Asia-Pacific, a telecom provider in Turkey, and a major media company in Asia. Additional reconnaissance referenced targets in South Korea, Egypt, Vietnam, Kenya, and a medical equipment manufacturer.
## Custom Tooling: AI-Generated Reconnaissance Framework
### CHECKER2 (Go)
A Docker-based orchestrator for parallel VPN scanning and target processing. Logs showed processing of 2,516 targets across 106 countries in parallel batches. Each container follows the flow: ingest stolen VPN config → attempt connection → scan internal network → pass results to ARXON for LLM analysis.
### ARXON MCP Server (Python)
A custom Model Context Protocol (MCP) server — no public references found, likely purpose-built. ARXON serves dual roles: 1. **Analysis platform**: Ingests per-target recon data, calls DeepSeek to generate attack plans, stores results in a persistent knowledge base that grows with each target. 2. **Attack toolkit**: Contains scripts to directly modify victim infrastructure — batch SSH-based FortiGate VPN account creation, user provisioning, automated Domain Admin credential validation.
### AI-Generated Reconnaissance Tools (Go/Python)
Custom post-VPN reconnaissance tools with clear AI-generation indicators: redundant comments restating function names, simplistic architecture with disproportionate investment in formatting over functionality, naive JSON parsing via string matching rather than proper deserialization, compatibility shims for language built-ins with empty documentation stubs. The tools automate: ingesting target networks from VPN routing tables, classifying networks by size, running gogo port scanner for service discovery, identifying SMB hosts and domain controllers, and Nuclei vulnerability scanning against HTTP services.
## Post-Exploitation Methodology
### Domain Compromise Operational documentation (in Russian) detailed the use of Meterpreter with the mimikatz module to perform DCSync attacks against domain controllers, extracting NTLM password hashes from Active Directory. In confirmed compromises, the attacker obtained complete domain credential databases. In at least one case, the Domain Administrator used a plaintext password extracted from FortiGate configuration through password reuse.
### Lateral Movement Following domain compromise: pass-the-hash and pass-the-ticket attacks against additional infrastructure, NTLM relay attacks using standard poisoning tools (ntlmrelayx.py from Impacket), and remote command execution on Windows hosts.
### Backup Infrastructure Targeting The actor specifically targeted Veeam Backup & Replication servers using DecryptVeeamPasswords.ps1 (PowerShell credential extraction), compiled decryption tools, and exploitation attempts leveraging CVE-2023-27532 (Veeam information disclosure) and CVE-2024-40711 (Veeam RCE). This is a classic pre-ransomware indicator — destroying backup infrastructure before encryption.
### Limited Exploitation Success A critical finding: the actor largely failed when attempting to exploit anything beyond straightforward automated attack paths. Their own documentation records repeated failures against patched services, closed ports, and non-matching OS versions. When encountering hardened environments, they simply moved on to softer targets.
## AI as Force Multiplier
### Multi-Model Operational Workflow - **DeepSeek**: Used for generating structured attack plans from reconnaissance data, target triage, and tactical analysis. - **Claude/Claude Code**: Used for vulnerability assessments during live intrusions, tool development, attack planning, and was configured to autonomously execute offensive tools (Impacket, Metasploit, hashcat) without human approval per command. - At least two distinct commercial LLM providers used in complementary roles.
### AI-Generated Attack Planning AI generated comprehensive attack methodologies with step-by-step exploitation instructions, expected success rates, time estimates, and prioritized task trees. Plans referenced academic research on offensive AI agents. The AI produced technically accurate command sequences, but the actor struggled to adapt when conditions differed from the plan.
### Operational Evolution The operation evolved over 8+ weeks: - **December 2025**: Used HexStrike (open-source offensive MCP framework). Claude Code pre-approved to execute Impacket, Metasploit, hashcat with hardcoded victim credentials. - **February 2026**: Graduated to custom ARXON MCP server — fully automated exploitation system with persistent knowledge base.
## Staging Infrastructure
The misconfigured staging server at 212.11.64.250:9999 (AS4264, Global-Data System IT Corporation, Zurich, Switzerland) exposed 1,402 files across 139 subdirectories, including: - Stolen FortiGate configuration backups - Active Directory mapping data (BloodHound collection data) - Credential dumps and NTLM hashes - Vulnerability assessment reports (Claude Code-generated) - CVE exploit code - Nuclei scanning templates - Veeam credential extraction tools - claude-0/ and claude/ directories with 200+ files (task outputs, session diffs, cached prompt states) - ARXON MCP server source code - CHECKER2 orchestrator - deploy_output.log showing 102MB archive of FortiGate configs organized by country - Russian-language operational documentation
A second staging server at 185.196.11.225 (Kali Linux) was used for automated scanning, receiving the 102MB config archive and launching parallel processing.
## Threat Actor Assessment (Amazon)
- **Motivation**: Financially motivated (pre-ransomware TTPs) - **Language**: Russian-speaking (operational documentation in Russian) - **Skill Level**: Low-to-medium baseline, significantly augmented by AI - **AI Dependency**: Extensive — all operational phases - **Operational Scale**: 600+ devices, 55+ countries, 5 weeks - **Post-Exploitation Depth**: Shallow — fails against hardened targets, moves on - **Operational Security**: Inadequate — unencrypted staging server with victim data
## Remediation
### Immediate Actions - Audit all FortiGate management interface exposure — disable internet-facing management ports - Enable MFA on all FortiGate administrative and VPN accounts - Rotate all SSL-VPN and administrative credentials on FortiGate devices - Check for unauthorized VPN accounts created via SSH (ARXON capability) - Review FortiGate configuration backups for signs of extraction - Audit Active Directory for DCSync indicators and NTLM relay artifacts
### Long-Term Defenses - Network segmentation — isolate management interfaces from internet - Implement certificate-based VPN authentication - Deploy Veeam backup hardening (immutable backups, MFA) - Monitor for AI-generated tooling patterns in endpoint logs - Audit SSH sessions to FortiGate appliances for bulk modifications - Deploy detection for CHECKER2/ARXON tool artifacts
MITRE ATT&CK techniques by tactic:
credential-access
T1003.001 LSASS Memory; T1003.006 DCSync; T1110.003 Password Spraying; T1552.001 Credentials In Files; T1557.001 Name Resolution Poisoning and SMB Relay; T1558 Steal or Forge Kerberos Tickets
discovery
T1016 System Network Configuration Discovery; T1018 Remote System Discovery; T1046 Network Service Discovery; T1069.002 Domain Groups; T1087.002 Domain Account
lateral-movement
T1021.002 SMB/Windows Admin Shares
exfiltration
T1041 Exfiltration Over C2 Channel
execution
T1059.001 PowerShell; T1059.004 Unix Shell; T1059.006 Python
defense-evasion
T1070 Indicator Removal; T1078 Valid Accounts; T1078.001 Default Accounts; T1078.002 Domain Accounts; T1550.002 Pass the Hash; T1550.003 Pass the Ticket
persistence
T1098 Account Manipulation; T1133 External Remote Services
command-and-control
T1105 Ingress Tool Transfer; T1219 Remote Access Tools
collection
T1213 Data from Information Repositories; T1602.002 Network Device Configuration Dump
impact
resource-development
Remediation
Patches
- Patch Veeam Backup & Replication for CVE-2023-27532 and CVE-2024-40711
- Patch QNAP NAS devices for CVE-2019-7192
- Update FortiGate firmware to latest stable release
- Enable FortiGate FGFM management plane hardening
Immediate actions
- Audit FortiGate management interface exposure — disable internet-facing management on ports 443, 8443, 10443, 4443
- Enable MFA on all FortiGate administrative and VPN accounts immediately
- Rotate all SSL-VPN and administrative credentials on FortiGate devices
- Check for unauthorized VPN user accounts created via SSH (ARXON batch provisioning)
- Review FortiGate configuration backups for signs of unauthorized extraction
- Audit Active Directory for DCSync indicators (Event ID 4662 with DS-Replication-Get-Changes)
- Check Veeam Backup servers for exploitation attempts (CVE-2023-27532, CVE-2024-40711)
- Block known staging IPs: 212.11.64.250, 185.196.11.225
Workarounds
- Restrict FortiGate management to internal/VPN-only access
- Disable recoverable password storage in FortiGate SSL-VPN config
- Enable SMB signing on all domain controllers and servers
- Implement geofencing on VPN access if organization has limited geographic footprint
Longer-term hardening
- Network segmentation — isolate management interfaces from internet access
- Implement certificate-based VPN authentication instead of password-only
- Deploy immutable backups and MFA on Veeam Backup infrastructure
- Monitor for AI-generated tooling patterns (redundant comments, naive parsing, formatting over functionality)
- Audit all SSH sessions to FortiGate appliances for bulk configuration changes
- Deploy EDR with behavioral detection for Impacket, Metasploit, mimikatz
- Implement NTLM relay protections (SMB signing, Extended Protection for Authentication)
- Regular credential audits — ensure VPN passwords differ from AD credentials
Affected and fixed versions
- Fortinet — FortiGate
Vulnerable versions: All versions with exposed management interfaces and weak credentials
Fixed in: N/A — configuration hardening required, not a software vulnerability - Veeam — Backup & Replication
Vulnerable versions: Versions vulnerable to CVE-2023-27532; Versions vulnerable to CVE-2024-40711
Fixed in: 12.1.2.172 and later (CVE-2024-40711); 12.0.0.1420 P20230718 and later (CVE-2023-27532) - QNAP — NAS
Vulnerable versions: Versions vulnerable to CVE-2019-7192
Fixed in: QTS 4.3.6.0923 and later - Microsoft — Active Directory
Vulnerable versions: All versions — targeted via DCSync with stolen credentials
Fixed in: N/A — credential-based attack
Timeline
- First exposure of staging server locked-ioc-8 detected by Hunt.io Attack Capture — contained HexStrike MCP framework, BloodHound data, Claude Code settings with hardcoded victim credentials from Asian media company. Source: https://cyberandramen.net/2026/02/21/llms-in-the-kill-chain-inside-a-custom-mcp-targeting-fortigate-devices-across-continents/
- Campaign begins — actor starts mass brute-force scanning of FortiGate management interfaces across ports 443, 8443, 10443, 4443. CHECKER2 Docker orchestrator deployed for parallel VPN scanning. Source: https://aws.amazon.com/blogs/security/ai-augmented-threat-actor-accesses-fortigate-devices-at-scale/
- Claude Code-generated vulnerability assessment report dated Feb 1, 2026 found on staging server, documenting active intrusion against Asia-Pacific industrial gas company — confirmed 400ms RTT to target. Impacket ntlmrelayx.py actively running. Source: Cyber and Ramen research
- Campaign activity window closes — 600+ FortiGate devices compromised across 55+ countries in 38 days. ARXON MCP evolved from HexStrike to fully automated exploitation system. Source: Amazon Threat Intelligence
- Amazon AWS Security Blog publishes research by CJ Moses (CISO). Cyber and Ramen publishes independent technical analysis of ARXON MCP and CHECKER2. BleepingComputer reports story. CronUp researcher German Fernandez identifies separate AI-generated FortiWeb tooling. Sources: AWS, Cyber and Ramen, BleepingComputer
- Threadlinqs Intelligence Platform publishes TL-2026-0131 with full MITRE mapping, detection coverage, and simulation documentation.
- As of 2026-05-29, the discrete 5-week campaign (Jan 11-Feb 18, 2026) has concluded, but the Russian-speaking actor remains unattributed with no arrest, takedown, or successor reported, and the AI-augmented TTPs plus exposed-FortiGate attack surface persist. The core vector is config-hardening, not a patchable CVE; secondary tool CVE-2024-40711 (Veeam) stays in CISA KEV under active ransomware exploitation, warranting continued monitoring rather than closure.
Sources cited
- NVD — CVE-2023-27532: Veeam Information Disclosure NVD (NIST) · Advisory · 2023-03-10
- NVD — CVE-2024-40711: Veeam RCE NVD (NIST) · Advisory · 2024-09-07
- NVD — CVE-2019-7192: QNAP RCE NVD (NIST) · Advisory · 2019-12-05
- Amazon AWS Security Blog: AI-Augmented Threat Actor Accesses FortiGate Devices at Scale aws.amazon.com · Research · 2026-02-21
- Cyber and Ramen: LLMs in the Kill Chain — Inside a Custom MCP Targeting FortiGate Devices Across Continents cyberandramen.net · Research · 2026-02-21
- VirusTotal: DecryptVeeamPasswords.ps1 from staging server virustotal.com · Research · 2026-02-21
- Hunt.io: Attack Capture — Staging Server Analysis hunt.io · Research · 2026-02-21
- Veeam KB4424: CVE-2023-27532 Advisory veeam.com · Vendor report · 2023-03-07
- Veeam KB4649: CVE-2024-40711 Advisory veeam.com · Vendor report · 2024-09-04
- HexStrike: Open-source Offensive Security MCP Framework hexstrike.com · Code repository · 2025-12-19
- BleepingComputer: AI-assisted hacker breached 600 Fortinet firewalls in 5 weeks BleepingComputer · News · 2026-02-21
- Google: Hackers Abusing Gemini AI for All Attack Stages BleepingComputer · News · 2026-02-15
Other reports on CVE-2023-27532
- Qilin Ransomware-as-a-Service Group: Leadership Attribution (Haise / Rjun / Alexbigboss44) and Most Active Affiliate Program
- ThreatsDay Roundup: AI API Key Leaks (OpenCode Console), PingCAP Source Code Theft, Zeabur Breach, NVIDIA AIStore CVE-2026-65162, Signal CVE-2026-113380, Laravel CVE-2026-114509
- #StopRansomware: SafePay Ransomware (CISA AA26-293A) — Exploitation of Cisco ASA/FTD, Veeam B&R and SonicWall SonicOS
- Gentlemen Ransomware (Storm-2697): Double-Extortion RaaS with Fortinet Exploitation, BYOVD EDR Killers and GPO-Based Mass Deployment
- Healthcare Sector Ransomware Targeting and Expanding Attack Surface (Qilin, Akira, DragonForce, LockBit, The Gentlemen)
- Qilin Ransomware Suspect Arrested in Japan, Extradited to Germany; Qilin Affiliate Exploited Check Point VPN Zero-Day CVE-2026-50751
Previous and next in apt
Drafted by the Threadlinqs research pipeline from 12 sources; published ; last updated —.