Skip to navigation

Threat reportVulnerabilityTL-2026-0094

n8n Multi-CVE Vulnerability Cascade — Expression Sandbox Escape, Pyodide RCE, Arbitrary File Write, Command Injection (9 CVEs, 2× CVSS 10.0)

criticalACTIVE

n8n workflow automation platform critical multi-CVE vulnerability cascade — 10+ CVEs disclosed Dec 2025–Feb 2026 including expression sandbox escape to RCE (CVE-2025-68613, CVE-2026-25049), Pyodide Python sandbox escape (CVE-2025-68668, CVSS 9.9), arbitrary file write to RCE (CVE-2026-21877, CVSS 10.0), unauthenticated file access via webhook (CVE-2026-21858, CVSS 10.0), Merge node file write RCE (CVE-2026-25056), Git node command injection (CVE-2026-25053), and file access TOCTOU leading to full account takeover (CVE-2026-25052).

Key facts

IdentifiersTL-2026-0094, CVE-2025-68613, CVE-2026-25049, CVE-2025-68668, CVE-2026-21877, CVE-2026-21858, CVE-2026-25056, CVE-2026-25053, CVE-2026-25052, CVE-2026-25115
SeverityCRITICAL
Exploitation statusLinked CVE exploited in the wild; Listed in CISA KEV: CVE-2025-68613 (added )
EPSSCVE-2025-68613: 98.9% (99.9th percentile); CVE-2026-21858: 78.1% (99.5th percentile); CVE-2025-68668: 13.1% (96.2th percentile)
Weaknesses (CWE)CWE-913, CWE-693, CWE-434, CWE-20, CWE-78, CWE-367
MITRE ATT&CK techniquesT1005 Data from Local System, T1053.003 Cron, T1059.004 Unix Shell, T1059.006 Python, T1059.007 JavaScript, T1068 Exploitation for Privilege Escalation, T1071.001 Web Protocols, T1082 System Information Discovery, T1083 File and Directory Discovery, T1098 Account Manipulation, T1190 Exploit Public-Facing Application, T1203 Exploitation for Client Execution and 13 more in the attack flow
MotivationUNKNOWN
Target sectorsTechnology, Enterprise, Financial Services, Healthcare, Manufacturing, Government
Target regionsGlobal
Detection rules9 (1 shown free)
Sources12 cited sources

Detection coverage

As of 2026-02-04, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-0094 across Splunk SPL, Microsoft KQL and Sigma, covering 32 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 Process Spawning Unexpected Child Processes — Confirmed Sandbox Escape RCE

Detects the n8n Node.js process spawning unexpected child processes (bash, sh, python3, curl, wget, nc, socat, perl) indicating confirmed sandbox escape exploitation. n8n should only spawn node (workers), git (Git node), and chromium (HTTP Request/Screenshot nodes). Any other child process is confirmed RCE — expression sandbox escape (CVE-2025-68613), Pyodide escape (CVE-2025-68668), command injection (CVE-2026-25053), or file write exploitation (CVE-2026-21877).

Splunk SPL

index=linux sourcetype=auditd type=EXECVE | eval n8n_parent=if(match(ppid_exe, "(?i)(node|n8n)") AND match(ppid_cmdline, "(?i)n8n"), 1, 0) | where n8n_parent=1 | eval child_risk=case(match(exe, "(?i)(/bash|/sh|/dash)$"), "CRITICAL_SHELL", match(exe, "(?i)(/python3?|/python)$"), "CRITICAL_PYTHON", match(exe, "(?i)(/curl|/wget)$"), "HIGH_DOWNLOAD", match(exe, "(?i)(/nc|/ncat|/socat)$"), "CRITICAL_REVERSE_SHELL", match(exe, "(?i)(/perl|/ruby|/php)$"), "HIGH_INTERPRETER", match(exe, "(?i)(/node|/git|/chromium)$"), "EXPECTED", 1=1, "UNKNOWN") | where child_risk!="EXPECTED" | stats count values(exe) as child_commands values(child_risk) as risk_levels latest(cmdline) as last_command by host | sort -count

Known false positives:

  • n8n Execute Command node legitimately running shell commands (should be restricted to trusted workflows only)
  • n8n community nodes that spawn helper processes

Tuning notes: CRITICAL_SHELL and CRITICAL_REVERSE_SHELL are ZERO false positive in properly configured n8n deployments — immediate incident response. Requires auditd or Falco configured to monitor n8n process children. If Execute Command node is used legitimately, whitelist specific expected commands. Any nc/ncat/socat = active reverse shell attempt.

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)
  • n8n Version Compliance — Unpatched Instances Vulnerable to Multi-CVE Cascade · unlock with TL Creds
  • n8n Webhook Path Traversal — Pre-Authentication File Access (CVE-2026-21858 CVSS 10.0) · unlock with TL Creds
  • n8n File Writes Outside Data Directory — Arbitrary Write Exploitation · unlock with TL Creds
  • n8n Code Node Executing Dangerous Imports — Sandbox Escape via Python/JavaScript · unlock with TL Creds
  • n8n Webhook Path Traversal Request — Pre-Authentication Attack (CVE-2026-21858) · unlock with TL Creds
  • n8n Process Creating Persistence Artifacts — Post-RCE Cron Jobs, SSH Keys, Webshells · unlock with TL Creds
  • n8n Expression Sandbox Escape — Dangerous Prototype Chain and Process Access Patterns · unlock with TL Creds
  • n8n Credential Bulk Access — Post-Exploitation Credential Theft Detection · unlock with TL Creds

Indicators of compromise

As of 2026-02-04, Threadlinqs Intelligence tracks 32 indicators of compromise for TL-2026-0094.

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

Technical analysis

Workflow automation platforms have code execution capabilities by design, making sandbox escapes catastrophic — authenticated users with workflow creation privileges achieve complete host server RCE.

n8n is a popular open-source workflow automation platform with 50,000+ GitHub stars and widespread enterprise deployment for business process automation, API integration, and AI agent orchestration. Between December 2025 and February 2026, n8n disclosed 10+ critical security vulnerabilities representing a systematic assault on the platform's security model. The primary vulnerability CVE-2025-68613 allows authenticated users with workflow creation privileges to escape the JavaScript expression sandbox through insufficient runtime isolation, achieving arbitrary code execution on the host server. CVE-2026-25049 represents additional expression escape exploits discovered after the initial fix. CVE-2025-68668 (CVSS 9.9) demonstrates that the Pyodide Python Code Node sandbox can also be escaped for RCE. Two CVEs scored CVSS 10.0: CVE-2026-21877 (arbitrary file write to RCE) and CVE-2026-21858 (unauthenticated file access via webhook). Additional critical vectors include CVE-2026-25056 (Merge node arbitrary file write to RCE), CVE-2026-25053 (Git node OS command injection), CVE-2026-25052 (file access TOCTOU race condition enabling arbitrary file read and full account takeover), and CVE-2026-25115 (additional Python sandbox escape). The breadth of vulnerabilities reveals a fundamental architectural challenge: workflow automation platforms grant users code execution capabilities by design through expression evaluation, code nodes, and system integrations — making every sandbox boundary a critical security surface. Public PoC exploits are available and active scanning has been observed targeting internet-exposed n8n instances.

MITRE ATT&CK techniques by tactic:

collection

T1005 Data from Local System

execution

T1053.003 Cron; T1059.004 Unix Shell; T1059.006 Python; T1059.007 JavaScript; T1203 Exploitation for Client Execution; T1204.002 Malicious File

privilege-escalation

T1068 Exploitation for Privilege Escalation; T1611 Escape to Host

command-and-control

T1071.001 Web Protocols

discovery

T1082 System Information Discovery; T1083 File and Directory Discovery

persistence

T1098 Account Manipulation; T1505.003 Web Shell

initial-access

T1190 Exploit Public-Facing Application

lateral-movement

T1210 Exploitation of Remote Services

defense-evasion

T1211 Exploitation for Stealth

impact

T1486 Data Encrypted for Impact; T1496 Resource Hijacking

credential-access

T1528 Steal Application Access Token; T1552.001 Credentials In Files; T1555 Credentials from Password Stores

exfiltration

T1567 Exfiltration Over Web Service

resource-development

T1583 Acquire Infrastructure

defense-impairment

T1685 Disable or Modify Tools

Remediation

Patches

  • n8n 2.5.2 (latest stable addressing all CVEs)
  • n8n 1.123.18 (LTS branch addressing all CVEs)
  • n8n 2.4.8 (Python sandbox escape fix)
  • n8n 2.0.0 (Pyodide sandbox escape fix)

Immediate actions

  • Update n8n to version 2.5.2 or later immediately — this version addresses all disclosed CVEs
  • Audit all n8n users with workflow creation privileges — each is a potential RCE vector
  • Restrict n8n instance access to authenticated, trusted users only — never expose to internet without authentication
  • Review all active workflows for suspicious expression patterns, Code node payloads, and Git node configurations
  • Enable n8n audit logging to detect exploitation attempts

Workarounds

  • Disable Code node and Python Code node for untrusted users
  • Disable Git node integration if not required
  • Restrict workflow creation to administrator-only
  • Block outbound network from n8n container to limit post-exploitation

Longer-term hardening

  • Deploy n8n in Docker containers with minimal host filesystem access and restricted network egress
  • Implement least-privilege access control — separate workflow viewers from workflow creators
  • Monitor n8n process for unexpected child processes, file system writes outside data directory, and outbound network connections
  • Subscribe to n8n security advisories on GitHub for ongoing vulnerability disclosure
  • Consider running n8n Code nodes in isolated execution environments (queue mode with separate workers)
  • Implement WAF rules for webhook endpoints to prevent unauthenticated exploitation

Timeline

  • Synacktiv publishes research on n8n multiple vulnerabilities (CVE-2023-27562, CVE-2023-27564) — SSRF and prototype pollution. Establishes n8n as a security research target. Source: https://www.synacktiv.com/sites/default/files/2023-05/Synacktiv-N8N-Multiple-Vulnerabilities_0.pdf
  • CVE-2025-68668 disclosed — Pyodide Python Code Node sandbox escape to RCE (CVSS 9.9). CWE-693 Protection Mechanism Failure. Patched in n8n 2.0.0. Source: GitHub Security Advisory
  • CVE-2026-21877 (CVSS 10.0, arbitrary file write to RCE) and CVE-2026-21858 (CVSS 10.0, unauthenticated file access via webhook) disclosed. Both represent critical pre-authentication attack vectors. Patched in 1.121.0/1.121.3. Source: GitHub Security Advisory
  • Active scanning observed targeting internet-exposed n8n instances for newly disclosed vulnerabilities. Public PoC exploits available for expression sandbox escape. Source: GitHub Security Advisory
  • n8n releases patches across multiple version branches: 2.5.2, 1.123.17, 1.123.18, 1.123.10. Addresses all disclosed expression escape, file write, command injection, and sandbox escape vulnerabilities. Source: https://github.com/n8n-io/n8n/releases
  • Massive vulnerability disclosure wave: CVE-2025-68613 (expression injection RCE), CVE-2026-25049 (expression escape follow-up, CVSS 9.4), CVE-2026-25056 (Merge node file write RCE), CVE-2026-25053 (Git node command injection), CVE-2026-25052 (file access TOCTOU account takeover), CVE-2026-25115 (Python sandbox escape). Source: GitHub Security Advisories
  • As of 2026-05-29, this n8n multi-CVE cascade remains actively exploited: CVE-2025-68613 was added to CISA KEV (Mar 11, due Mar 25) with confirmed Zerobot botnet exploitation since mid-Jan and a public Metasploit module. Patches exist, but Intel 471/Censys/Cyera report 26K-100K still-exposed instances and the CVSS-10.0 unauth "Ni8mare" (CVE-2026-21858) has live PoCs and active scanning.

Sources cited

Other reports on CVE-2025-68613

Previous and next in vulnerability

Drafted by the Threadlinqs research pipeline from 12 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