Threat reportVulnerabilityTL-2026-0226
n8n Workflow Automation RCE via Expression Injection — CVE-2025-68613 (CVSS 9.9) Active Exploitation by Zerobot Botnet
Critical RCE vulnerability (CVSS 9.9) in n8n workflow automation platform versions 0.211.0 through 1.120.3 and 1.121.0, actively exploited in the wild by the Zerobot botnet to deploy Mirai-variant malware. Insufficient sandbox isolation in the expression evaluation engine allows authenticated attackers to escape the JavaScript sandbox via Node.js this context and execute arbitrary OS commands.
Key facts
| Identifiers | TL-2026-0226, CVE-2025-68613 |
|---|---|
| Also known as | GHSA-v98v-ff95-f3cp |
| Severity | CRITICAL |
| CVSS | 9.9 (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H) |
| Exploitation status | Linked CVE exploited in the wild; Listed in CISA KEV: CVE-2025-68613 (added ) |
| EPSS | CVE-2025-68613: 98.9% (99.9th percentile) |
| Affected products | n8n-io n8n |
| Weaknesses (CWE) | CWE-913 |
| Attribution | Zerobot Botnet (medium confidence) |
| Malware | zerobotv9 |
| Tools | Mirai/Zerobot |
| MITRE ATT&CK techniques | T1005 Data from Local System, T1027 Obfuscated Files or Information, T1041 Exfiltration Over C2 Channel, T1053 Scheduled Task/Job, T1059 Command and Scripting Interpreter, T1068 Exploitation for Privilege Escalation, T1070 Indicator Removal, T1071 Application Layer Protocol, T1082 System Information Discovery, T1083 File and Directory Discovery, T1087 Account Discovery, T1105 Ingress Tool Transfer and 7 more in the attack flow |
| Motivation | FINANCIAL |
| Target sectors | technology, financial, healthcare, government, manufacturing, education, retail |
| Target regions | North America, Europe, Asia-Pacific, Global |
| Detection rules | 9 (1 shown free) |
| Sources | 13 cited sources |
Detection coverage
As of 2026-03-14, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-0226 across Splunk SPL, Microsoft KQL and Sigma, covering 23 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: n8n Expression Injection RCE — Node.js Spawning Shell Process (SPL)
Detects a Node.js process (n8n server) spawning shell interpreters (bash, sh, cmd, powershell), which is the primary indicator of CVE-2025-68613 sandbox escape. When the expression injection payload executes, the n8n node process calls child_process.execSync(), causing a shell child process. This is highly anomalous for n8n in production.
- Languages: Splunk SPL, Microsoft KQL, Sigma
- ATT&CK: T1190, T1059.004, T1059.007, T1203
- Data sources: sysmon, windows_security, linux_auditd, edr
- Severity: critical
- Confidence: high
Splunk SPL
index=sysmon sourcetype="XmlWinEventLog:Microsoft-Windows-Sysmon/Operational" EventCode=1
| search (ParentImage="*\\node.exe" OR ParentImage="*/node")
| search (Image="*\\cmd.exe" OR Image="*\\powershell.exe" OR Image="*\\pwsh.exe" OR Image="*/bash" OR Image="*/sh" OR Image="*/dash")
| eval risk_score=case(
match(CommandLine, "(?i)child_process|execSync|mainModule|require"), 100,
match(CommandLine, "(?i)whoami|id|uname|cat /etc"), 90,
match(CommandLine, "(?i)wget|curl|chmod"), 85,
1=1, 70
)
| where risk_score >= 70
| table _time host ParentImage Image CommandLine User risk_score
| sort - risk_scoreMicrosoft KQL
DeviceProcessEvents
| where Timestamp > ago(24h)
| where (InitiatingProcessFileName =~ "node" or InitiatingProcessFileName =~ "node.exe")
| where (FileName in~ ("bash", "sh", "dash", "cmd.exe", "powershell.exe", "pwsh.exe"))
| extend RiskScore = case(
ProcessCommandLine has_any ("child_process", "execSync", "mainModule", "require"), 100,
ProcessCommandLine has_any ("whoami", "id", "uname", "cat /etc"), 90,
ProcessCommandLine has_any ("wget", "curl", "chmod"), 85,
true, 70
)
| where RiskScore >= 70
| project Timestamp, DeviceName, InitiatingProcessFileName, FileName, ProcessCommandLine, AccountName, RiskScore
| sort by RiskScore descSigma
title: n8n Expression Injection RCE — Node.js Spawning Shell Process
id: 8a3f7e12-b4c5-4d6a-9e1f-2a3b4c5d6e7f
status: experimental
description: Detects Node.js (n8n) spawning shell interpreters indicating CVE-2025-68613 expression sandbox escape via this.process.mainModule.require('child_process').execSync()
references:
- https://nvd.nist.gov/vuln/detail/CVE-2025-68613
- https://github.com/n8n-io/n8n/security/advisories/GHSA-v98v-ff95-f3cp
author: Threadlinqs Team
date: 2026-03-14
tags:
- attack.initial_access
- attack.t1190
- attack.execution
- attack.t1059.004
- attack.t1059.007
- cve.2025.68613
logsource:
category: process_creation
product: linux
detection:
selection_parent:
ParentImage|endswith: '/node'
selection_child:
Image|endswith:
- '/bash'
- '/sh'
- '/dash'
condition: selection_parent and selection_child
falsepositives:
- Legitimate Node.js applications with intentional shell execution (e.g., build tools, task runners)
- n8n Execute Command nodes used by administrators for approved automation tasks
level: criticalKnown false positives:
- Legitimate Node.js build tools or task runners that spawn shell processes
- n8n Execute Command nodes used by trusted administrators for approved automation
Tuning notes: Exclude known legitimate Node.js applications that spawn shells (e.g., npm scripts, webpack). Add allowlist for specific ParentCommandLine values containing known safe n8n workflow IDs. Adjust risk_score thresholds based on environment baseline.
- n8n REST API Expression Injection — Malicious Workflow Creation (KQL) · unlock with TL Creds
- n8n Post-Exploitation — Credential Harvesting & Persistence (KQL) · unlock with TL Creds
- Zerobot DNS Resolution & C2 Beacon Pattern — Non-Standard Port (KQL) · unlock with TL Creds
- n8n Expression Injection — Windows Node.exe Spawning PowerShell/CMD (Sigma) · unlock with TL Creds
- Executable Binary Dropped in Temp Directory After Download — Zerobot Pattern (Sigma) · unlock with TL Creds
- Zerobot zerobotv9 Binary Execution & File Hash Match (Sigma) · unlock with TL Creds
- Zerobot [indicator] Downloader & zerobotv9 Payload Staging (SPL) · unlock with TL Creds
- Zerobot C2 Communication — [indicator] and Infrastructure IPs (SPL) · unlock with TL Creds
Indicators of compromise
As of 2026-03-14, Threadlinqs Intelligence tracks 23 indicators of compromise for TL-2026-0226.
Sample indicators (3 of 23)
- IP address:
144.172.100[.]228 - Domain:
0bot[.]qzz[.]io - File hash:
c8e8b627398ece071a3a148d6f38e46763dc534f9bfd967ebc8ac3479540111f
Defanged, TLP:CLEAR. Full set of 23 indicators: unlock with TL Creds.
Technical analysis
CISA added CVE-2025-68613 to the KEV catalog on March 11, 2026 with a federal patch deadline of March 25, 2026. Over 24,700 unpatched instances remain exposed globally.
CVE-2025-68613 is a critical Remote Code Execution vulnerability in n8n, a popular open-source workflow automation platform used by thousands of organizations for business process automation, data integration, and AI workflow orchestration.
The vulnerability resides in n8n's server-side expression evaluation engine, which processes JavaScript expressions wrapped in {{ }} delimiters within workflow node parameters. The core flaw is that the execution sandbox fails to properly isolate the evaluation context from the Node.js runtime. Specifically, the JavaScript this keyword within evaluated expressions exposes the Node.js global context, granting access to the process object. An attacker can leverage this to call process.mainModule.require('child_process').execSync() to execute arbitrary OS-level commands with the privileges of the n8n process.
The exploit chain is straightforward: 1. An authenticated user (no elevated privileges required) creates or modifies a workflow via the REST API (POST /rest/workflows) 2. A malicious expression is injected into a node parameter (typically using the n8n-nodes-base.set node type) 3. When the workflow is activated or executed, the expression evaluator processes the payload 4. The JavaScript escapes the sandbox via this.process.mainModule.require() 5. The child_process module is loaded and execSync() executes arbitrary commands
The canonical exploit payload is: {{ (function(){ return this.process.mainModule.require('child_process').execSync('COMMAND').toString() })() }}
Akamai's Security Intelligence and Response Team (SIRT) identified active exploitation by the Zerobot botnet beginning in mid-January 2026, with campaign activity traced back to December 2025. The botnet exploits CVE-2025-68613 to deploy a Mirai-based malware variant called zerobotv9, which supports multiple architectures (x86, MIPS, ARM, PPC). The initial infection vector uses a shell script (tol.sh) downloaded from 144.172.100.228 that fetches and executes architecture-specific zerobotv9 binaries. The C2 infrastructure operates via 0bot.qzz[.]io.
Shadowserver Foundation data from February 2026 shows 24,700+ unpatched n8n instances exposed globally, with 12,300+ in North America and 7,800+ in Europe. Censys reported 103,476 potentially vulnerable instances as of December 22, 2025. The EPSS score is 76.93% (99th percentile), indicating extremely high probability of exploitation.
The vulnerability was patched in December 2025 with versions 1.120.4, 1.121.1, and 1.122.0, which introduce additional safeguards to restrict expression evaluation and prevent sandbox escapes. A related vulnerability, CVE-2026-27577 (CVSS 9.4), was subsequently discovered by Pillar Security in the same expression evaluation system.
Organizations running self-hosted n8n instances must patch immediately or implement strict access controls to limit workflow creation and editing to trusted users only.
MITRE ATT&CK techniques by tactic:
collection
defense-evasion
T1027 Obfuscated Files or Information; T1070 Indicator Removal; T1140 Deobfuscate/Decode Files or Information
exfiltration
T1041 Exfiltration Over C2 Channel
execution
T1053 Scheduled Task/Job; T1059 Command and Scripting Interpreter; T1203 Exploitation for Client Execution
privilege-escalation
T1068 Exploitation for Privilege Escalation
command-and-control
T1071 Application Layer Protocol; T1105 Ingress Tool Transfer; T1571 Non-Standard Port
discovery
T1082 System Information Discovery; T1083 File and Directory Discovery; T1087 Account Discovery
initial-access
T1190 Exploit Public-Facing Application
impact
persistence
T1505 Server Software Component
credential-access
Remediation
Patches
- n8n v1.120.4 — backport fix for expression sandbox hardening
- n8n v1.121.1 — backport fix for expression sandbox hardening
- n8n v1.122.0 — primary fix release with enhanced expression evaluation safeguards
Immediate actions
- Upgrade n8n to version 1.120.4, 1.121.1, or 1.122.0+ immediately
- Block IOC IP addresses at network perimeter (144.172.100.228, 140.233.190.96, 172.86.123.179, 216.126.227.101, 103.59.160.237)
- Block C2 domain 0bot.qzz.io and malware distribution domains andro.notemacro.com, pivot.notemacro.com at DNS and proxy
- Restrict workflow creation and editing permissions to trusted users only
- Disable public registration on n8n instances
- Audit existing workflows for suspicious expressions containing process.mainModule or child_process references
Workarounds
- Restrict workflow creation/editing to fully trusted administrators only
- Disable n8n REST API access from untrusted networks
- Place n8n behind a reverse proxy with authentication and IP allowlisting
- Run n8n with minimal OS privileges and restricted filesystem access
- Monitor for expressions containing process, require, child_process, execSync in workflow definitions
Longer-term hardening
- Deploy n8n in hardened environments with restricted OS privileges — never run as root
- Implement network segmentation to isolate n8n instances from critical infrastructure
- Restrict outbound network connections from n8n to only required endpoints
- Enable comprehensive audit logging for workflow creation and modification
- Deploy EDR/XDR with behavioral detection for Node.js process spawning child processes
- Implement API key rotation and credential management for n8n integrations
- Monitor for CVE-2026-27577 (related expression evaluation flaw) and apply patches when available
Affected and fixed versions
- n8n-io — n8n
Vulnerable versions: 0.211.0 through 1.120.3; 1.121.0
Fixed in: 1.120.4; 1.121.1; 1.122.0
Timeline
- CVE-2025-68613 published in NVD; n8n releases security patches v1.120.4, v1.121.1, v1.122.0 with expression sandbox hardening
- Censys identifies 103,476 potentially vulnerable n8n instances exposed to the internet; public PoC exploit code published on GitHub
- Spike in CVE-2025-68613 exploitation detected around Christmas period as attackers target holiday-period reduced staffing
- Resecurity publishes detailed technical analysis of CVE-2025-68613 including sandbox escape mechanism and exploit chain
- Akamai SIRT detects Zerobot botnet actively exploiting CVE-2025-68613 in global honeypot network; Mirai-variant zerobotv9 payloads observed
- Shadowserver Foundation reports 24,700+ unpatched n8n instances still exposed globally — 12,300+ in North America, 7,800+ in Europe
- Akamai SIRT publishes comprehensive analysis of Zerobot botnet campaign exploiting CVE-2025-68613 and CVE-2025-7544 with full IOCs
- CISA adds CVE-2025-68613 to Known Exploited Vulnerabilities catalog confirming active exploitation; FCEB agencies given March 25, 2026 deadline to patch per BOD 22-01
- Threadlinqs Intelligence documents threat with full MITRE mappings, IOCs, and Zerobot botnet correlation for platform publication
- As of 2026-05-29, CVE-2025-68613 (n8n expression-injection RCE, CVSS 9.9) remains actively exploited: it sits in CISA KEV (added Mar 11), is exploited by the operating Zerobot botnet (Akamai/Intel 471), with 24,700+ unpatched internet-exposed instances and public PoCs. Patches exist, but mass exploitation continues and follow-on n8n CVEs extend the same flaw class.
Sources cited
- CISA KEV Addition — CVE-2025-68613 CISA · Government advisory · 2026-03-11
- NVD — CVE-2025-68613 NVD (NIST) · Advisory · 2025-12-19
- GitHub Security Advisory GHSA-v98v-ff95-f3cp GitHub · Advisory · 2025-12-19
- CVE-2025-68613: Critical n8n RCE & Server Compromise — Orca Security orca.security · Research · 2026-03-11
- CVE-2025-68613: RCE via Expression Injection in n8n — Resecurity resecurity.com · Research · 2025-12-26
- Akamai SIRT — Zerobot Malware Targets n8n Automation Platform akamai.com · Research · 2026-02-28
- Akamai SIRT Identifies Zerobot Botnet Exploiting n8n — ThreatIntelReport threatintelreport.com · Research · 2026-02-28
- CVE-2025-68613 Authenticated Expression-Injection RCE in n8n — Penligent penligent.ai · Research · 2026-03-11
- PoC Exploit for CVE-2025-68613 — Expression Injection RCE in n8n GitHub · Proof of concept · 2025-12-22
- Nuclei Detection Template for CVE-2025-68613 GitHub · Code repository · 2025-12-22
- n8n CVE-2025-68613 RCE Exploitation: A Detailed Guide — SecureLayer7 blog.securelayer7.net · Blog · 2025-12-21
- CVE-2025-68613: Critical RCE in n8n — SOCRadar socradar.io · Blog · 2026-03-11
- CISA Flags Actively Exploited n8n RCE Bug — 24,700 Instances Exposed The Hacker News · News · 2026-03-11
Other reports on CVE-2025-68613
- Kaspersky ICS CERT: APT and Financial Attacks on Industrial Organizations in Q2 2026 (DAEMON Tools Supply Chain, Iran-Linked PLC Attacks, ZionSiphon, INC Ransomware)
- CVE-2026-34486: Apache Tomcat EncryptInterceptor Bypass Actively Exploited in SnowLight and AI-Assisted Campaigns
- Seedworm (MuddyWater) Iranian MOIS APT Campaign Targeting U.S. Critical Infrastructure with Dindoor and Fakeset Backdoors
Previous and next in vulnerability
Drafted by the Threadlinqs research pipeline from 13 sources; published ; last updated —.