CVE-2025-68613: n8n: security flaw enables exploitation

HIGHACTIVELY EXPLOITEDPoC AVAILABLENUCLEI TEMPLATECISA: ACT

Published December 19, 2025

CISO Take

CVE-2025-68613 is a CVSS 8.8 RCE in n8n, actively exploited in the wild (CISA KEV, Zerobot malware campaign) and requires only low-privilege authentication — patch to 1.120.4/1.121.1/1.122.0 immediately. If n8n is part of your AI agent infrastructure, assume credential compromise: rotate all API keys, LLM tokens, and integration secrets stored in the platform. Any n8n instance accessible from the internet with multi-user access must be treated as breached until patched.

What is the risk?

Critical operational risk for organizations using n8n in AI agent pipelines or automation workflows. CVSS 8.8 with network-accessible attack vector and low privilege requirement means any authenticated user — including trial accounts or compromised internal accounts — can achieve full instance compromise. CISA KEV inclusion and a documented Zerobot malware campaign confirm active exploitation at scale. Risk is amplified in AI contexts because n8n instances typically hold credentials for LLM APIs (OpenAI, Anthropic), vector databases, cloud storage, and SaaS integrations — a single compromise cascades across the entire AI stack. Exploitation complexity is low; no specialized AI/ML knowledge required.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
n8n	npm	—	No patch
209.2KOpenSSF 7.116 dependentsPushed 5d ago44% patched~19d to patchFull package profile →

Do you use n8n? You're affected.

How severe is it?

CVSS 3.1

8.8 / 10

EPSS

99.0%

chance of exploitation in 30 days

Higher than 100% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Actively Exploited

CISA KEV

Sophistication

Trivial

Exploitation Confidence

high

✓CISA KEV (active exploitation confirmed) — Mar 2026

✓CISA SSVC: Active exploitation

✓Weaponized Metasploit module (exploit/multi/http/n8n_workflow_expression_rce)

○Nuclei detection template available

○EPSS exploit prediction: 99%

Composite signal derived from CISA KEV, VulnCheck KEV, CISA SSVC, EPSS, Metasploit, Exploit-DB, trickest/cve, Nuclei templates, and inthewild.io exploitation reports.

What is the attack surface?

AVNetwork

ACLow

PRLow

UINone

SUnchanged

CHigh

IHigh

AHigh

What should I do?

7 steps

PATCH IMMEDIATELY
Upgrade to n8n 1.120.4, 1.121.1, or 1.122.0. This is the only complete fix.
If patching is blocked: restrict workflow creation/editing to fully trusted users only and isolate n8n behind VPN or internal network — remove all public internet exposure.
Rotate all credentials stored in n8n (LLM API keys, database passwords, OAuth tokens, webhook secrets) regardless of whether compromise is confirmed.
Audit workflow modification history for unauthorized changes, particularly any new HTTP Request nodes, Execute Command nodes, or newly created credentials.
Deploy n8n in a hardened container with restricted OS privileges (non-root, read-only filesystem where possible, egress firewall rules).
Detection: Monitor for unusual n8n process spawning child processes (bash, sh, cmd), unexpected outbound connections from the n8n host, and new workflow creation by non-admin users. Alert on access to /etc/passwd, SSH key directories, or credential files from the n8n process.
Check n8n logs for expression payloads containing process execution patterns (child_process, exec, spawn, eval).

What does CISA's SSVC say?

DecisionAct

Exploitationactive

AutomatableNo

Technical Impacttotal

Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.

How is it classified?

Code ExecutionAuth BypassData ExtractionAgentFrameworkPluginAML.T0012- Valid Accounts AML.T0049- Exploit Public-Facing Application AML.T0050- Command and Scripting Interpreter AML.T0053- AI Agent Tool Invocation AML.T0072- Reverse Shell AML.T0081- Modify AI Agent Configuration AML.T0083- Credentials from AI Agent Configuration AML.T0105- Escape to Host AML.T0108- AI Agent

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Art. 15-Accuracy, robustness and cybersecurityArt. 9-Risk management systemArticle 15-Accuracy, robustness and cybersecurityArticle 9-Risk management system

ISO 42001

A.6.1.4-AI risk assessmentA.6.2-AI system roles and responsibilitiesA.9.2-Information security in AI system development and operationA.9.7-Information security in AI system operations

NIST AI RMF

GOVERN 1.2-Policies and processes related to AI risk are establishedGOVERN 6.1-Policies and procedures for AI risk managementMANAGE 2.2-Mechanisms to sustain AI risk management are planned

OWASP LLM Top 10

LLM05:2025-Improper Output HandlingLLM06-Sensitive Information DisclosureLLM06:2025-Excessive AgencyLLM08-Excessive Agency

Frequently Asked Questions

What is CVE-2025-68613?

Is CVE-2025-68613 actively exploited?

Yes, CVE-2025-68613 is confirmed actively exploited and listed in CISA Known Exploited Vulnerabilities catalog since Wed Mar 11 2026 00:00:00 GMT+0000 (Coordinated Universal Time).

How to fix CVE-2025-68613?

1. PATCH IMMEDIATELY: Upgrade to n8n 1.120.4, 1.121.1, or 1.122.0. This is the only complete fix. 2. If patching is blocked: restrict workflow creation/editing to fully trusted users only and isolate n8n behind VPN or internal network — remove all public internet exposure. 3. Rotate all credentials stored in n8n (LLM API keys, database passwords, OAuth tokens, webhook secrets) regardless of whether compromise is confirmed. 4. Audit workflow modification history for unauthorized changes, particularly any new HTTP Request nodes, Execute Command nodes, or newly created credentials. 5. Deploy n8n in a hardened container with restricted OS privileges (non-root, read-only filesystem where possible, egress firewall rules). 6. Detection: Monitor for unusual n8n process spawning child processes (bash, sh, cmd), unexpected outbound connections from the n8n host, and new workflow creation by non-admin users. Alert on access to /etc/passwd, SSH key directories, or credential files from the n8n process. 7. Check n8n logs for expression payloads containing process execution patterns (child_process, exec, spawn, eval).

What systems are affected by CVE-2025-68613?

This vulnerability affects the following AI/ML architecture patterns: agent frameworks, LLM integration platforms, RAG pipelines, AI workflow automation, model serving.

What is the CVSS score for CVE-2025-68613?

CVE-2025-68613 has a CVSS v3.1 base score of 8.8 (HIGH). The EPSS exploitation probability is 98.99%.

What is the AI security impact?

Affected AI Architectures

agent frameworksLLM integration platformsRAG pipelinesAI workflow automationmodel serving

MITRE ATLAS Techniques

AML.T0012Valid Accounts

AML.T0049Exploit Public-Facing Application

AML.T0050Command and Scripting Interpreter

AML.T0053AI Agent Tool Invocation

AML.T0072Reverse Shell

AML.T0081Modify AI Agent Configuration

AML.T0083Credentials from AI Agent Configuration

AML.T0105Escape to Host

AML.T0108AI Agent

Compliance Controls Affected

EU AI Act: Art. 15, Art. 9, Article 15, Article 9

ISO 42001: A.6.1.4, A.6.2, A.9.2, A.9.7

NIST AI RMF: GOVERN 1.2, GOVERN 6.1, MANAGE 2.2

OWASP LLM Top 10: LLM05:2025, LLM06, LLM06:2025, LLM08

What are the technical details?

Original Advisory

n8n is an open source workflow automation platform. Versions starting with 0.211.0 and prior to 1.120.4, 1.121.1, and 1.122.0 contain a critical Remote Code Execution (RCE) vulnerability in their workflow expression evaluation system. Under certain conditions, expressions supplied by authenticated users during workflow configuration may be evaluated in an execution context that is not sufficiently isolated from the underlying runtime. An authenticated attacker could abuse this behavior to execute arbitrary code with the privileges of the n8n process. Successful exploitation may lead to full compromise of the affected instance, including unauthorized access to sensitive data, modification of workflows, and execution of system-level operations. This issue has been fixed in versions 1.120.4, 1.121.1, and 1.122.0. Users are strongly advised to upgrade to a patched version, which introduces additional safeguards to restrict expression evaluation. If upgrading is not immediately possible, administrators should consider the following temporary mitigations: Limit workflow creation and editing permissions to fully trusted users only; and/or deploy n8n in a hardened environment with restricted operating system privileges and network access to reduce the impact of potential exploitation. These workarounds do not fully eliminate the risk and should only be used as short-term measures.

Exploitation Scenario

Attacker obtains low-privilege n8n credentials via phishing, credential stuffing, or purchasing them on darknet markets (n8n is widely deployed as a SaaS-style internal tool). They create a new workflow containing a malicious expression — e.g., in a Function node or expression field — that abuses CWE-913 (improper control of dynamically-managed code resources) to break out of the expression sandbox and execute OS commands. With the privileges of the n8n process, they execute a reverse shell, establish persistence, and immediately target n8n's encrypted credential store. They exfiltrate all stored API keys (OpenAI, Anthropic, AWS, Slack, GitHub, etc.), then modify existing AI agent workflows to silently forward all processed data to an attacker-controlled endpoint. In the Zerobot campaign context, they also enroll the host in a botnet for cryptomining or DDoS operations. The entire attack chain from initial access to full AI pipeline compromise takes under 10 minutes.

Weaknesses (CWE)

CWE-913Improper Control of Dynamically-Managed Code ResourcesPrimary CWE-913Improper Control of Dynamically-Managed Code Resources

CWE-913 — Improper Control of Dynamically-Managed Code Resources: The product does not properly restrict reading from or writing to dynamically-managed code resources such as variables, objects, classes, attributes, functions, or executable instructions or statements.

[Implementation] For any externally-influenced input, check the input against an allowlist of acceptable values.
[Implementation, Architecture and Design] Refactor the code so that it does not need to be dynamically managed.

Source: MITRE CWE corpus.