Threat reportThreat IntelligenceTL-2026-6152
GTIG warns fully-automated agentic AI attacks imminent: exploit_agent, CAIRN, ARTEX abuse and SOMBERMEME via modified Claude Code (MIDNIGHT NEPTUNE / UNC6836)
Google Threat Intelligence Group's AI Threat Tracker (2026-11-16) reports that threat actors fully automated vulnerability discovery and exploitation without a human in the loop for the first time, and estimates low-sophistication attacks needing little human interaction within ~3 months and sophisticated fully autonomous attacks within ~6 months.
Key facts
| Identifiers | TL-2026-6152, CVE-2026-81578, CVE-2026-82078, CVE-2026-109618, CVE-2026-109619, CVE-2026-40217, CVE-2021-42278, CVE-2021-42287 |
|---|---|
| Also known as | Agentic Threat Activity Crosses a Significant Threshold, GTIG AI Threat Tracker November 2026 |
| Severity | HIGH |
| First disclosed | |
| Exploitation status | Linked CVE exploited in the wild; Listed in CISA KEV: CVE-2026-81578 (added ), CVE-2026-82078 (added ), CVE-2026-109618 (added ), CVE-2026-109619 (added ), CVE-2021-42278 (added ), CVE-2021-42287 (added ) |
| EPSS | CVE-2026-109618: 99.9% (99.9th percentile); CVE-2026-109619: 99.9% (99.9th percentile); CVE-2026-81578: 85.5% (99.7th percentile) |
| Affected products | Anthropic Claude Code (abused via .claude/settings.json configuration); Various Internet-facing web applications (WordPress Spectra plugin v2.19.25, PaperCut, PHP) targeted by agentic exploitation |
| Attribution | MIDNIGHT NEPTUNE (low confidence) |
| Nation-state nexus | North Korea |
| Malware | PHP, SOMBERMEME |
| Tools | ARTEX, ARTEX PRO, CAIRN, CyberStrikeAI, Hermes, HexStrike AI, PentAGI, Strix, exploit_agent |
| MITRE ATT&CK techniques | T1003.004 OS Credential Dumping: LSA Secrets, T1003.006 OS Credential Dumping: DCSync, T1021.006 Remote Services: Windows Remote Management, T1036 Masquerading, T1036.005 Masquerading: Match Legitimate Resource Name or Location, T1059 Command and Scripting Interpreter, T1059.001 Command and Scripting Interpreter: PowerShell, T1068 Exploitation for Privilege Escalation, T1071.001 Application Layer Protocol: Web Protocols, T1078.004 Valid Accounts: Cloud Accounts, T1090 Proxy, T1098 Account Manipulation and 30 more in the attack flow |
| Motivation | FINANCIAL |
| Target sectors | aerospace, cryptocurrency, news - media, entertainment, education, finance, government administration, technology, defense, health, retail |
| Target regions | North America, Europe, france, Asia, united states of america |
| Detection rules | 9 (1 shown free) |
| Sources | 5 cited sources |
Detection coverage
As of 2026-11-17, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-6152 across Splunk SPL, Microsoft KQL and Sigma, covering 59 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: SOMBERMEME Delivery - Claude Code Launch Spawning Download Child Process (TL-2026-6152)
Detects the supply-chain delivery pattern attributed to MIDNIGHT NEPTUNE (suspected DPRK) where a modified .claude/settings.json in an internal repository causes Claude Code startup to download the SOMBERMEME backdoor. The search looks for process-creation events where claude (or node running claude) is the parent and the child command line contains download primitives (URL, DownloadString, DownloadFile, Invoke-WebRequest, certutil urlcache, bitsadmin). Behaviour-first: it does not depend on any published hash or domain, because GTIG published none.
- Languages: Splunk SPL, Microsoft KQL, Sigma
- ATT&CK: T1195, T1204, T1036
- Data sources: sysmon, microsoft_defender_for_endpoint
- Severity: high
- Confidence: high
Splunk SPL
index=sysmon sourcetype="XmlWinEventLog:Microsoft-Windows-Sysmon/Operational" EventCode=1 earliest=-24h (ParentImage="*claude.exe" OR ParentImage="*/claude" OR (ParentImage="*node*" AND ParentCommandLine="*claude*")) (CommandLine="*http://*" OR CommandLine="*https://*" OR CommandLine="*DownloadString*" OR CommandLine="*DownloadFile*" OR CommandLine="*Invoke-WebRequest*" OR CommandLine="*urlcache*" OR CommandLine="*bitsadmin*") | stats count min(_time) as first_seen max(_time) as last_seen values(CommandLine) as child_commands values(ParentCommandLine) as claude_cmdline by host, User, ParentImage, Image | eval first_seen=strftime(first_seen,"%Y-%m-%d %H:%M:%S"), last_seen=strftime(last_seen,"%Y-%m-%d %H:%M:%S") | table first_seen last_seen host User ParentImage Image child_commands claude_cmdline countMicrosoft KQL
DeviceProcessEvents
| where Timestamp > ago(24h)
| where InitiatingProcessFileName in~ ("claude.exe", "claude") or (InitiatingProcessFileName in~ ("node.exe", "node") and InitiatingProcessCommandLine has "claude")
| where FileName in~ ("curl.exe", "curl", "wget", "powershell.exe", "pwsh.exe", "pwsh", "cmd.exe", "bash", "sh", "certutil.exe", "bitsadmin.exe", "mshta.exe")
| where ProcessCommandLine matches regex @"(?i)(https?://|DownloadString|DownloadFile|Invoke-WebRequest|\biwr\b|\birm\b|urlcache|bitsadmin|curl |wget )"
| project Timestamp, DeviceName, AccountName, InitiatingProcessFileName, InitiatingProcessCommandLine, FileName, ProcessCommandLine, SHA256Sigma
title: Claude Code Spawning Download Child Process - SOMBERMEME Delivery Pattern
id: 7c1e4a52-3b8d-4f6a-9e21-5d0a8c3b7f14
status: experimental
description: Detects Claude Code (or node running Claude Code) launching a child process that downloads content, the delivery pattern GTIG attributes to MIDNIGHT NEPTUNE abusing a modified .claude/settings.json to fetch the SOMBERMEME backdoor.
references:
- https://cloud.google.com/blog/topics/threat-intelligence/agentic-threat-activity-crosses-a-significant-threshold
- https://www.bleepingcomputer.com/news/security/google-warns-that-fully-automated-ai-attacks-are-coming-soon/
author: Threadlinqs Team
date: 2026-11-16
tags:
- attack.initial-access
- attack.t1195
- attack.execution
- attack.t1204
- attack.defense-evasion
- attack.t1036
logsource:
category: process_creation
product: windows
detection:
selection_parent_claude:
ParentImage|endswith: '\claude.exe'
selection_parent_node:
ParentImage|endswith: '\node.exe'
ParentCommandLine|contains: 'claude'
selection_download:
CommandLine|contains:
- 'http://'
- 'https://'
- 'DownloadString'
- 'DownloadFile'
- 'Invoke-WebRequest'
- 'urlcache'
- 'bitsadmin'
condition: (selection_parent_claude or selection_parent_node) and selection_download
falsepositives:
- Developers asking Claude Code to run curl or package installs that fetch from URLs
- Legitimate Claude Code hooks that call internal tooling over HTTP
level: highKnown false positives:
- Developers legitimately instructing Claude Code to run curl, wget or package managers that fetch from URLs
- Sanctioned Claude Code hooks in .claude/settings.json that call internal tooling over HTTP
Tuning notes: Baseline approved hook commands per repository and allow-list known internal download hosts. Pair with SIG-001 (settings.json modification) to correlate the config change with the first launch. Extend ParentImage patterns if Claude Code is installed under a different binary name on your fleet.
- Agentic Web Exploitation - High-Velocity Path Probing and Spectra [indicator] Access · unlock with TL Creds
- Password Spraying Against Entra ID - Many Accounts Failing From One IP · unlock with TL Creds
- Entra ID Cloud Application Permission Abuse and Credential Modification Burst · unlock with TL Creds
- Claude Code settings.json Written by Script Interpreter or Download Utility (SOMBERMEME Supply-Chain Vector) · unlock with TL Creds
- AI Offensive Framework Execution - ARTEX, CAIRN, PentAGI, CyberStrikeAI, HexStrike AI, exploit_agent · unlock with TL Creds
- Community-Observed DPRK-Tagged Infrastructure and File Hash Watchlist (Low Confidence) · unlock with TL Creds
- Web Shell Activity - Web Server Process Spawning Shell or Recon Utilities (exploit_agent / ARTEX) · unlock with TL Creds
- LLMjacking - AWS Bedrock Model Enumeration, Burst Invocation or Invocation Logging Tampering · unlock with TL Creds
Indicators of compromise
As of 2026-11-17, Threadlinqs Intelligence tracks 59 indicators of compromise for TL-2026-6152.
Sample indicators (3 of 59)
- IP address:
45.142.193[.]132 - Domain:
jet-night[.]com - File hash:
528cd4e69ecfa5191adbcf6ef28667bf
Defanged, TLP:CLEAR. Full set of 59 indicators: unlock with TL Creds.
Technical analysis
Observed activity includes the exploit_agent framework, abuse of CAIRN/ARTEX/PentAGI/CyberStrikeAI/HexStrike AI, and a suspected DPRK cluster (MIDNIGHT NEPTUNE, reported by BleepingComputer as UNC6836) delivering the SOMBERMEME backdoor via modified Claude Code settings.
GTIG's AI Threat Tracker, 'Agentic Threat Activity Crosses a Significant Threshold' (Blum, Vanderlee, Xia, Mainor; 2026-11-16), documents a shift from AI-assisted to agent-driven intrusion activity. GTIG states that in the past quarter it observed threat actors fully automate vulnerability discovery and exploitation without a human in the loop for the first time, and that these frameworks sharply shorten the time defenders have to patch.
exploit_agent is a multi-agent vulnerability research and weaponization framework found on an attacker's exposed staging server. It integrates with frontier and local LLMs, performs continuous differential analysis of web application updates, drafts exploits and executes them without human intervention. The staging server held reconnaissance output, stolen victim data, brute-forcing scripts and webshell installation scripts, including scripts that remove other attackers' webshells. Victims were French media, entertainment and education websites. GTIG also describes an agentic framework analyzing Spectra (Ultimate Addons for Gutenberg) v2.19.25 (file v21925.php), a WordPress plugin with an insecure access-control check.
Open-source offensive frameworks are being abused: CAIRN (multi-agent pentest framework using an OODA loop and shared blackboard; observed in intrusions against the U.S. aerospace sector), ARTEX (Go-based autonomous pentest framework weaponized by multiple actors since August 2026; an FBI FLASH advisory was released 2026-11-13 and Google SecOps has deployed detection rulepacks), PentAGI (a customized version used for recon and scanning of organizations in Asia from late August to early October 2026), CyberStrikeAI (LLM-native framework with a Go backend acting as C2/orchestration) and HexStrike AI (MCP-integrated harness connecting LLMs to 150+ offensive tools via 12+ sub-agents), the latter two seen in experimentation. Per the FBI advisory as reported by BleepingComputer, ARTEX agents perform recon, exploitation, privilege escalation and data theft, including password spraying, cloud application permission abuse, credential modification and web shell deployment, reaching privileged access in under three hours. Separate Gambit reporting describes a financially motivated operator chaining Strix (scanning), Cairn (exploitation) and Hermes (orchestration) to compromise at least 27 companies and inject skimmers into 119+ sites.
GTIG attributes SOMBERMEME delivery to MIDNIGHT NEPTUNE, a suspected DPRK cluster that combines LLM-enabled social engineering with abuse of AI tools as a supply-chain vector: the actor modified .claude/settings.json in internal repositories so that launching Claude Code triggers download of the SOMBERMEME backdoor. Targets are cryptocurrency organizations in North America and Europe. BleepingComputer reports the cluster as UNC6836 with a suspected link to Midnight Neptune; the GTIG article itself does not use the UNC6836 identifier. GTIG additionally reports LLMjacking (theft of LLM credentials and cloud resources resold via LLM Mixers) with single-day charges over $100,000 and cumulative losses up to $1M, and LLM-assisted exploitation of PHP flaws (CVE-2026-109618/-109619; patches 2026-10-21) and PaperCut zero-days (CVE-2026-81578, CVE-2026-82078) exploited four days after disclosure. No hashes, IPs or domains are published in the sources reviewed; IOCs here are tool, malware, filename and entity indicators. Attribution of SOMBERMEME to DPRK is 'suspected'.
MITRE ATT&CK techniques by tactic:
Credential Access
T1003.004 OS Credential Dumping: LSA Secrets; T1003.006 OS Credential Dumping: DCSync; T1110.003 Brute Force: Password Spraying; T1528 Steal Application Access Token; T1539 Steal Web Session Cookie; T1552 Unsecured Credentials; T1552.001 Unsecured Credentials: Credentials In Files; T1558.003 Steal or Forge Kerberos Tickets: Kerberoasting; T1649 Steal or Forge Authentication Certificates
Lateral Movement
T1021.006 Remote Services: Windows Remote Management; T1550.002 Use Alternate Authentication Material: Pass the Hash
Defense Evasion
T1036 Masquerading; T1036.005 Masquerading: Match Legitimate Resource Name or Location; T1562.008 Impair Defenses: Disable or Modify Cloud Logs
Execution
T1059 Command and Scripting Interpreter; T1059.001 Command and Scripting Interpreter: PowerShell; T1204 User Execution; T1204.004 User Execution: Malicious Copy and Paste
Privilege Escalation
T1068 Exploitation for Privilege Escalation
Command and Control
T1071.001 Application Layer Protocol: Web Protocols; T1090 Proxy; T1105 Ingress Tool Transfer; T1572 Protocol Tunneling
Initial Access
T1078.004 Valid Accounts: Cloud Accounts; T1190 Exploit Public-Facing Application; T1195 Supply Chain Compromise; T1195.002 Supply Chain Compromise: Compromise Software Supply Chain; T1566 Phishing; T1566.003 Phishing: Spearphishing via Service
Persistence
T1098 Account Manipulation; T1098.001 Account Manipulation: Additional Cloud Credentials; T1136.003 Create Account: Cloud Account; T1505.003 Server Software Component: Web Shell
Impact
T1496 Resource Hijacking; T1496.004 Resource Hijacking: Cloud Service Hijacking
Exfiltration
T1567 Exfiltration Over Web Service
Resource Development
T1583 Acquire Infrastructure; T1587.004 Develop Capabilities: Exploits; T1588.007 Obtain Capabilities: Artificial Intelligence; T1650 Acquire Access
Reconnaissance
T1595 Active Scanning; T1595.002 Active Scanning: Vulnerability Scanning
Remediation
Patches
- Apply PHP patches released 2026-10-21 (CVE-2026-109618, CVE-2026-109619)
- Apply PaperCut fixes for CVE-2026-81578 and CVE-2026-82078
Immediate actions
- Review and monitor changes to AI agent configuration files such as .claude/settings.json in internal repositories as a supply-chain vector
- Deploy Google SecOps ARTEX rulepacks and other ARTEX behavioral/volumetric detections (e.g. MalwareMustDie ARTEX signature pack)
- Rotate and revoke credentials and session tokens; keep secrets out of public repositories and CI environments
- Monitor for unexpected LLM service enablement, quota increase requests, GPU instance provisioning and bursts of LLM API calls (LLMjacking)
Workarounds
- Restrict internet exposure of web apps and review application permissions and authentication logs
- Report ARTEX-related activity to the FBI or IC3
Longer-term hardening
- Shorten patch windows because attackers can weaponize n-days by diffing patches
- Apply Mandiant hardening guidance and architectural defenses against machine-speed attacks
- Rate-limit sensitive APIs and restrict access to cloud administration interfaces
- Treat LLM-generated PoCs with skepticism and verify before acting
Affected and fixed versions
- Anthropic — Claude Code (abused via .claude/settings.json configuration)
- Various — Internet-facing web applications (WordPress Spectra plugin v2.19.25, PaperCut, PHP) targeted by agentic exploitation
Vulnerable versions: Spectra (Ultimate Addons for Gutenberg) 2.19.25
Timeline
Last updated · 3 updates · revalidated 3× · latest source
- Obsidian Security published the LiteLLM chain (CVE-2026-47101, CVE-2026-47102, CVE-2026-40217); GTIG later observed CVE-2026-40217 used in LLMjacking intrusions.
- ARTEX autonomous pentest framework released on GitHub by developer Autumn-27 (July 2026; exact day not stated).
- Since at least August 2026, multiple threat actors weaponize ARTEX against organizations across several sectors (FBI); Mandiant also sees ARTEX, PentAGI and CAIRN in incident response from late August.
- Huntress observed exploitation of the PaperCut NG/MF chain (CVE-2026-81578 + CVE-2026-82078) as zero-days; LLM-assisted exploits appeared within a day of disclosure.
- GreyNoise-observed AI-orchestrated PaperCut campaign: empty workspace to first RCE in four hours, first domain admin about two hours later.
- PaperCut flaws added to CISA KEV (sources give 2026-08-31 or 2026-09-01); GreyNoise reported a campaign four days after disclosure.
- Campaign relaunched from a second host; 11 organizations compromised in 26 seconds and over 223 systems by end of day (GreyNoise).
- GreyNoise-observed AI-orchestrated PaperCut campaign window (2026-08-27 to 2026-09-02) ends after at least 440 instances at 395 organizations in 48 countries were targeted.
- GreyNoise publishes analysis: at least 440 PaperCut instances at 395 organizations in 48 countries, 12 with Domain Admin.
- Gambit reporting: financially motivated operator using Strix, Cairn and Hermes agents compromised 27+ companies and infected 119+ sites with skimmers; campaign ongoing.
- ARTEX developer Autumn-27 removes the GitHub repository; community forks and the commercial ARTEX PRO remain available.
- Wiz publishes 'ARTEX in the Wild' on Storm-0940 targeting of IT and software vendors; Microsoft's STORM-0940 report is reported the same day.
- Official PHP patches released for CVE-2026-109618 and CVE-2026-109619, after PoCs and patched code were published beforehand.
- GTIG/Mandiant publish 'ARTEX and Friends'; Google SecOps releases ARTEX rulepacks and Mandiant a hardening guide.
- GTIG/Mandiant publish the LLMjacking report covering LLM mixers, single-day fraud above $100K, up to $1M cumulative losses and CVE-2026-40217 abuse.
- Trojanized ARTEX Docker image (tomicdd/artex) served by 62 exposed consoles between 2026-11-02 and 2026-11-04 per Cloud Security Alliance; about 9,000 pulls.
- Censys reported 1,465 hosts running an exposed ARTEX console.
- FBI FLASH advisory warns multiple actors have weaponized ARTEX; includes IPs and C2 infrastructure details in the PDF.
- GTIG publishes AI Threat Tracker: first observed fully automated vulnerability discovery and exploitation without a human in the loop; projects low-human-interaction attacks in ~3 months and sophisticated fully autonomous attacks in ~6 months.
Update history
- 2026-11-17 — Autonomous AI Agent Frameworks (ARTEX, CAIRN, PentAGI, exploit_agent) Used to Exploit Vulnerabilities Without Human Intervention - GTIG AI Threat Tracker: What changed No severity or exploitability change (already HIGH / ACTIVE). Record enriched with additional supporting detail from the same GTIG tracker and its cited sources.
- 2026-11-16 — Google GTIG: AI Agents Automate Exploitation, Credential Theft (LLMjacking) and Cybercrime at Scale (ARTEX, CAIRN, PentAGI, SOMBERMEME): What changed No severity, exploitability or status change; the existing record is already HIGH / ACTIVE. The update is enrichment only. New indicators (26) Indicator values are withheld at this access level. New CVEs (0) None.
- 2026-11-16 — Threat Actors Use Autonomous AI Agent Frameworks to Reverse Engineer Patches and Run Exploits (GTIG): What changed No severity, exploitability, status or attribution field changes; HIGH / ACTIVE already reflect the evidence. Record is enriched additively. New indicators (6) Indicator values are withheld at this access level.
Sources cited
- GTIG AI Threat Tracker: Agentic Threat Activity Crosses a Significant Threshold Google Cloud · Research · 2026-11-16
- Google warns that fully-automated AI attacks are coming soon (BleepingComputer) BleepingComputer · News · 2026-11-16
- FBI warns ARTEX AI hacking tool is powering global cyberattacks (BleepingComputer) BleepingComputer · News · 2026-11-15
- Malicious AI agents steal 600K credit cards, infect 100+ sites with skimmers (BleepingComputer) BleepingComputer · News · 2026-09-22
- Google: LLMjacking attacks rack up $1 million in cloud AI charges (BleepingComputer) BleepingComputer · News · 2026-11-16
More from MIDNIGHT NEPTUNE
Other reports on CVE-2026-81578
- Cisco Secure FMC Exploitation (CVE-2026-20079, CVE-2026-20316) by Qilin Affiliate UAT-11988 and Sandworm-Linked UAT-11823, with ScreenConnect (CVE-2026-84869) and PaperCut (CVE-2026-81578/82078) Management-Layer Exploitation
- NCSC CTO Weekly Summary (week ending 2026-09-13): actively exploited Cisco FMC (CVE-2026-20079/20316), PaperCut NG/MF (CVE-2026-81578/82078), MikroTik RouterOS and FortiGate (CVE-2025-25249) flaws; NetScaler SAML bypass (CVE-2026-19490); BlueMoon exploit kit
- September 2026 Major Cyber Attacks: ShinyHunters FBIJobs/PeopleSoft (CVE-2026-35273), Clop Leak Site Grav CMS Hijack (CVE-2026-42608), PaperCut AI-Agent Campaign, BlueMoon Zero-Day Chain, NetScaler C2 (CVE-2026-88771)
- Abliterated and Resold AI Models Fueling Cybercrime Underground; ARTEX-Assisted Intrusions at South Korean Banks (Flare Quantitative Analysis)
- AI-Agent-Driven Exploitation of PaperCut NG/MF (CVE-2026-81578, CVE-2026-82078) Leads to Windows Credential Theft and Domain Admin Compromise
- September 2026 CVE Landscape: 94 high-impact vulnerabilities, NetScaler zero-days CVE-2026-88771/CVE-2026-88772, AI-agent-assisted exploitation
Other reports on CVE-2026-82078
Previous and next in threat intel
Drafted by the Threadlinqs research pipeline from 5 sources; published ; last updated .