Threat reportRansomwareTL-2026-3786
Ransomware at Machine Speed: GenAI Shrinks the Path After Initial Access
Zscaler ThreatLabz reports that ransomware operators use public-facing GenAI platforms to craft convincing phishing and social-engineering content, speed up discovery of sensitive data, and rapidly generate custom scripts and tools that evade signature-based detection. The article names no specific ransomware family, CVE, or IOC; supporting context comes from the Zscaler 2026 Ransomware Report and related RMM-abuse reporting. First disclosed .
Key facts
| Identifiers | TL-2026-3786 |
|---|---|
| Also known as | GenAI-assisted ransomware operations |
| Severity | MEDIUM |
| First disclosed | |
| Affected products | Various Enterprise endpoints, email and collaboration platforms (Microsoft Teams, Quick Assist) and RMM tooling abused by operators |
| Malware | AgendaCrypt, Python, inc ransom |
| Tools | ConnectWise ScreenConnect, Kaseya (VSA), LogMeIn Resolve, Quick Assist, ScreenConnect, Syspectr |
| MITRE ATT&CK techniques | T1005 Data from Local System, T1027 Obfuscated Files or Information, T1059.001 Command and Scripting Interpreter: PowerShell, T1059.006 Command and Scripting Interpreter: Python, T1059.007 Command and Scripting Interpreter: JavaScript, T1219 Remote Access Tools, T1566 Phishing, T1566.003 Phishing: Spearphishing via Service, T1567 Exfiltration Over Web Service, T1587.001 Develop Capabilities: Malware, T1588.007 Obtain Capabilities: Artificial Intelligence, T1598 Phishing for Information |
| Motivation | FINANCIAL |
| Target sectors | manufacturing, technology, freight-logistics, utilities |
| Target regions | North America, Europe |
| Detection rules | 9 (1 shown free) |
| Sources | 5 cited sources |
Detection coverage
As of 2026-10-16, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-3786 across Splunk SPL, Microsoft KQL and Sigma, covering 15 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: Unsanctioned RMM Client Execution (ScreenConnect, LogMeIn Resolve, Kaseya, Syspectr) — TL-2026-3786
Detects execution of commercial RMM clients (ConnectWise ScreenConnect, LogMeIn Resolve, Kaseya, O&O Syspectr) that GenAI-assisted ransomware operators deploy as a post-phish foothold (T1219; Zscaler ThreatLabz, Forescout SeasonalInvite). Behaviour-first: it keys on process image, original file name and the ScreenConnect guest-session command line, then summarises per host and user so unsanctioned installs stand out against the approved RMM inventory.
- Languages: Splunk SPL, Microsoft KQL, Sigma
- ATT&CK: T1219, T1566
- Data sources: sysmon, index=sysmon sourcetype=XmlWinEventLog:Microsoft-Windows-Sysmon/Operational, DeviceProcessEvents
- Severity: high
- Confidence: high
Splunk SPL
index=sysmon sourcetype="XmlWinEventLog:Microsoft-Windows-Sysmon/Operational" EventCode=1 (Image="*ScreenConnect.Client*" OR Image="*ScreenConnect.Windows*" OR Image="*LogMeInResolve*" OR Image="*Syspectr*" OR Image="*KaseyaEndpoint*" OR Image="*AgentMon.exe" OR OriginalFileName="ScreenConnect*" OR CommandLine="*e=Access&y=Guest*")
| eval rmm_tool=case(like(Image,"%ScreenConnect%") OR like(CommandLine,"%e=Access&y=Guest%"),"ConnectWise ScreenConnect", like(Image,"%LogMeIn%"),"LogMeIn Resolve", like(Image,"%Syspectr%"),"O&O Syspectr", like(Image,"%Kaseya%") OR like(Image,"%AgentMon%"),"Kaseya", true(),"other_rmm")
| stats count min(_time) as first_seen max(_time) as last_seen values(ParentImage) as parent_image values(CommandLine) as command_line by host, User, rmm_tool
| 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 rmm_tool parent_image command_line countMicrosoft KQL
DeviceProcessEvents
| where Timestamp > ago(24h)
| where FileName has_any ("ScreenConnect.ClientService.exe", "ScreenConnect.WindowsClient.exe", "LogMeInResolve", "Syspectr", "KaseyaEndpoint", "AgentMon.exe") or ProcessCommandLine has "e=Access&y=Guest"
| extend RmmTool = case(FileName has "ScreenConnect" or ProcessCommandLine has "e=Access&y=Guest", "ConnectWise ScreenConnect", FileName has "LogMeIn", "LogMeIn Resolve", FileName has "Syspectr", "O&O Syspectr", "Kaseya")
| summarize Executions=count(), FirstSeen=min(Timestamp), LastSeen=max(Timestamp), ParentProcesses=make_set(InitiatingProcessFileName, 5), CommandLines=make_set(ProcessCommandLine, 3) by DeviceName, AccountName, RmmTool
| project FirstSeen, LastSeen, DeviceName, AccountName, RmmTool, Executions, ParentProcesses, CommandLinesSigma
title: Unsanctioned RMM Client Execution (ScreenConnect, LogMeIn Resolve, Kaseya, Syspectr)
id: 3f6a1c52-8b4e-4d97-a1c3-5e2b7d9f0a11
status: experimental
description: Detects execution of commercial RMM clients abused by ransomware operators as a foothold after phishing or vishing (TL-2026-3786).
references:
- https://www.zscaler.com/blogs/security-research/ransomware-machine-speed-how-genai-shrinks-path-after-initial-access
- https://www.forescout.com/blog/seasonalinvite-new-phishing-campaign-abuses-ecards-and-rmm/
author: Threadlinqs Team
date: 2026-10-16
tags:
- attack.command_and_control
- attack.t1219
- attack.initial_access
- attack.t1566
logsource:
category: process_creation
product: windows
detection:
selection_image:
Image|contains:
- 'ScreenConnect.ClientService'
- 'ScreenConnect.WindowsClient'
- 'LogMeInResolve'
- 'Syspectr'
- 'KaseyaEndpoint'
- 'AgentMon.exe'
selection_cmd:
CommandLine|contains: 'e=Access&y=Guest'
condition: 1 of selection_*
falsepositives:
- Sanctioned IT helpdesk or MSP use of the same RMM products
level: mediumKnown false positives:
- Sanctioned IT helpdesk or MSP deployments of ScreenConnect, LogMeIn Resolve, Kaseya or Syspectr
- Software-deployment tooling that pushes RMM agents during onboarding
Tuning notes: Exclude hosts and ParentImage values belonging to your approved RMM inventory (for example your own ScreenConnect instance host name in CommandLine). Alert only on first-seen rmm_tool per host if the estate runs sanctioned RMM.
- Fake Teams Installer MSI Execution (Teamsetup.msi Lure) · unlock with TL Creds
- Script Interpreters (Python, Node, WScript) Running From User-Writable Paths · unlock with TL Creds
- Network Connections to SeasonalInvite Infrastructure and RMM Relays · unlock with TL Creds
- Encoded PowerShell With Download Cradle (Generated Script Loader) · unlock with TL Creds
- PowerShell Script Block Recursive Document Discovery and Staging · unlock with TL Creds
- DNS Resolution of SeasonalInvite Domains and Scripted R2 Staging Lookups · unlock with TL Creds
- Quick Assist Session Followed by Shell Activity (Vishing Foothold) · unlock with TL Creds
- Data Exfiltration to Web Services and SeasonalInvite Staging Hosts via Proxy · unlock with TL Creds
Indicators of compromise
As of 2026-10-16, Threadlinqs Intelligence tracks 15 indicators of compromise for TL-2026-3786.
Sample indicators (3 of 15)
- IP address:
64.34.84[.]93 - Domain:
lanero[.]vu - IP address:
46.4.120[.]162
Defanged, TLP:CLEAR. Full set of 15 indicators: unlock with TL Creds.
Technical analysis
Zscaler ThreatLabz (Phil Hagen, published 2026-10-16) describes how ransomware operators apply generative AI to the stages between initial access and extortion. Operators use free or stolen LLM accounts on public-facing GenAI platforms to produce polished phishing and social-engineering material, to accelerate discovery and classification of sensitive data on victim systems, and to generate custom scripts and tools quickly enough to evade signature-based detection. Named tradecraft in the article is phishing, deployment of Remote Monitoring and Management (RMM) tools for a foothold, data discovery and exfiltration, and evasion of malware detection. The article does not name a ransomware family, actor, CVE, or IOC, and gives no timing statistics.
The companion Zscaler ThreatLabz 2026 Ransomware Report (released 2026-09-30, analysis window April 2025 to March 2026) provides context: data theft rose 275% year over year to 896.2 TB, blockchain-traced ransomware payments totaled $328 million, the average payment rose 5.3% to $431,995, and 7,366 victims were listed on leak sites (down 3%). Qilin, Akira and INC Ransom together accounted for 34% of disclosed victims, 52 newly active groups were identified, and 62% of victims held manager-level titles or above. Manufacturing and technology were the most targeted sectors; freight and logistics (+725%) and utilities (+622%) grew fastest; the United States accounted for 50.7% of activity, followed by Canada (4.8%), Germany (4.3%) and the UK (4.1%). The report also notes abuse of Microsoft Teams and Quick Assist for social engineering, lateral movement and data theft, and expanding script-based tooling in JavaScript, PowerShell and Python. Extortion is shifting from encryption toward data theft.
Independent reporting corroborates LLM-assisted tradecraft around RMM abuse. Forescout's SeasonalInvite analysis (active January to at least late June 2026) found phishing pages with HTML comments and emoji patterns consistent with LLM-generated code, delivering ConnectWise ScreenConnect, LogMeIn Resolve, Kaseya and O&O Syspectr. That campaign is not attributed to the ransomware operators in the Zscaler article; it is recorded here as related context only. Severity is an analyst estimate because the source is trend-level and lacks specifics.
MITRE ATT&CK techniques by tactic:
Collection
Defense Evasion
T1027 Obfuscated Files or Information
Execution
T1059.001 Command and Scripting Interpreter: PowerShell; T1059.006 Command and Scripting Interpreter: Python; T1059.007 Command and Scripting Interpreter: JavaScript
Command and Control
Initial Access
T1566 Phishing; T1566.003 Phishing: Spearphishing via Service
Exfiltration
T1567 Exfiltration Over Web Service
Resource Development
T1587.001 Develop Capabilities: Malware; T1588.007 Obtain Capabilities: Artificial Intelligence
Reconnaissance
Remediation
Immediate actions
- Restrict and inventory RMM tooling; alert on installation or execution of unsanctioned RMM clients (ScreenConnect, LogMeIn Resolve, Kaseya, O&O Syspectr, Quick Assist)
- Restrict external Microsoft Teams contact and block or limit Quick Assist where not required
- Monitor and control access to public GenAI platforms and flag use of unsanctioned or personal LLM accounts from corporate endpoints
Workarounds
- Application control allow-listing for remote-access software
- Block web categories for newly registered or uncategorized domains and for file-sharing storage used for staging
Longer-term hardening
- Deploy behavior-based EDR and script-block logging (PowerShell, JavaScript, Python) rather than relying on signatures
- Apply DLP and egress monitoring to detect bulk data discovery and exfiltration
- Run phishing and vishing awareness training that assumes lures are grammatically flawless and personalized
- Prioritize protection of managers, executives and other privileged roles that are disproportionately targeted
- Adopt AI-assisted detection and triage on the defensive side, as the source recommends
Affected and fixed versions
- Various — Enterprise endpoints, email and collaboration platforms (Microsoft Teams, Quick Assist) and RMM tooling abused by operators
Timeline
- Start of the analysis period (April 2025 to March 2026) covered by the Zscaler ThreatLabz 2026 Ransomware Report
- Forescout reports the SeasonalInvite phishing campaign abusing commercial RMM tools active since January 2026 (related context; month-level date)
- End of the Zscaler 2026 Ransomware Report analysis period
- SeasonalInvite activity observed through at least late June 2026, with LLM-style code indicators in phishing pages (month-level date)
- Zscaler releases the 2026 Ransomware Report: data theft up 275% to 896.2 TB, GenAI used to accelerate operations, executives and managers targeted
- Zscaler ThreatLabz publishes 'Ransomware at Machine Speed' describing GenAI use for phishing, data discovery and custom script generation after initial access
Sources cited
- Ransomware at Machine Speed: How GenAI Shrinks the Path after Initial Access (Zscaler ThreatLabz) zscaler.com · Research · 2026-10-16
- New Zscaler Report Reveals AI-Assisted Attackers Move to Massive Data Theft and Executive Targeting zscaler.com · Vendor report · 2026-09-30
- SeasonalInvite: New Phishing Campaign Abuses eCards and RMM (Forescout) forescout.com · Research
- Daisy-Chaining Rogue RMM Tools: How Threat Actors Abuse Remote Management Software for Initial Access (Huntress) Huntress · Blog
- Why Generative AI Fuels Ransomware Attacks Through Language, Not Code (Bitdefender) bitdefender.com · Blog
Previous and next in ransomware
- Cactus Ransomware Exploiting Qlik Sense Servers (CVE-2023-41266, CVE-2023-41265 'ZeroQlik', CVE-2023-48365 'DoubleQlik') - Fox-IT Victim Identification Previous report
- DragonForce Ransomware Delivered via Compromised MSP SimpleHelp RMM (CVE-2024-57726/57727/57728) — Managed Service Supply-Chain Attack Next report
Further reading
Drafted by the Threadlinqs research pipeline from 5 sources; published ; last updated —.