CVE-2026-25053: n8n Command Injection enables RCE

CISO Take

n8n is widely deployed as an AI agent orchestration platform and this 9.9 CVE gives any authenticated user—including low-privilege accounts—arbitrary OS command execution and file read on the host. Patch immediately to 1.123.10 or 2.5.0; treat all API keys, AI service credentials, and secrets stored on any n8n host as potentially compromised. If patching is not immediately possible, restrict workflow creation and modification to trusted administrators only.

What is the risk?

Extremely high. CVSS 9.9 with Scope:Changed means blast radius extends beyond the n8n process—attackers can pivot to the underlying host and all connected services. Low complexity + low privilege + no user interaction makes this trivially exploitable by any authenticated user. n8n instances are frequently internet-exposed with privileged access to AI APIs (OpenAI, Anthropic), databases, and internal services through automation workflows, dramatically amplifying the effective impact.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
n8n	npm	—	No patch
193.4K OpenSSF 6.6 Pushed 2d ago 55% patched ~7d to patch Full package profile →

Do you use n8n? You're affected.

How severe is it?

CVSS 3.1

9.9 / 10

EPSS

0.6%

chance of exploitation in 30 days

Higher than 43% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

No known exploitation

Sophistication

Trivial

What is the attack surface?

AV Network

AC Low

PR Low

UI None

S Changed

C High

I High

A High

What should I do?

7 steps

PATCH IMMEDIATELY

Upgrade to n8n 1.123.10 (v1 branch) or 2.5.0 (v2 branch)—no workaround substitutes for patching.
ROTATE ALL CREDENTIALS

Assume any secret accessible from the n8n host (AI API keys, DB passwords, SSH keys, Stripe/webhook secrets, .env vars) is compromised if patch was delayed.
RESTRICT PERMISSIONS

Limit workflow create/modify rights to trusted admins until patched.
NETWORK ISOLATION

Ensure n8n is not publicly exposed; place behind VPN or strict IP allowlist.
AUDIT LOGS

Review workflow creation and modification events for unauthorized or suspicious Git node usage in the period before patching.
HARDEN CONTAINER

Run n8n with minimal host filesystem access, read-only mounts, and dropped capabilities.
DETECTION

Alert on unexpected outbound connections from the n8n host and anomalous child process spawning (e.g., bash, sh, curl spawned by the n8n process).

What does CISA's SSVC say?

Decision Track

Exploitation none

Automatable No

Technical Impact total

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

How is it classified?

Code Execution Data Extraction Auth Bypass Agent Framework Plugin AML.T0012 - Valid Accounts AML.T0037 - Data from Local System AML.T0049 - Exploit Public-Facing Application AML.T0050 - Command and Scripting Interpreter AML.T0053 - AI Agent Tool Invocation AML.T0055 - Unsecured Credentials AML.T0072 - Reverse Shell AML.T0083 - Credentials from AI Agent Configuration AML.T0086 - Exfiltration via AI Agent Tool Invocation

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Art.15 - Accuracy, robustness and cybersecurity Article 15 - Accuracy, Robustness and Cybersecurity

ISO 42001

A.6.2.6 - AI system security

NIST AI RMF

GOVERN-1.7 - AI Supply Chain Risk Management GOVERN-6.1 - Policies and procedures for AI risk management MANAGE-2.2 - Mechanisms are in place and tested to manage AI risks

OWASP LLM Top 10

LLM02 - Insecure Output Handling LLM06 - Excessive Agency LLM07 - Insecure Plugin Design LLM08 - Excessive Agency

Frequently Asked Questions

What is CVE-2026-25053?

n8n is widely deployed as an AI agent orchestration platform and this 9.9 CVE gives any authenticated user—including low-privilege accounts—arbitrary OS command execution and file read on the host. Patch immediately to 1.123.10 or 2.5.0; treat all API keys, AI service credentials, and secrets stored on any n8n host as potentially compromised. If patching is not immediately possible, restrict workflow creation and modification to trusted administrators only.

Is CVE-2026-25053 actively exploited?

No confirmed active exploitation of CVE-2026-25053 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-25053?

1. PATCH IMMEDIATELY: Upgrade to n8n 1.123.10 (v1 branch) or 2.5.0 (v2 branch)—no workaround substitutes for patching. 2. ROTATE ALL CREDENTIALS: Assume any secret accessible from the n8n host (AI API keys, DB passwords, SSH keys, Stripe/webhook secrets, .env vars) is compromised if patch was delayed. 3. RESTRICT PERMISSIONS: Limit workflow create/modify rights to trusted admins until patched. 4. NETWORK ISOLATION: Ensure n8n is not publicly exposed; place behind VPN or strict IP allowlist. 5. AUDIT LOGS: Review workflow creation and modification events for unauthorized or suspicious Git node usage in the period before patching. 6. HARDEN CONTAINER: Run n8n with minimal host filesystem access, read-only mounts, and dropped capabilities. 7. DETECTION: Alert on unexpected outbound connections from the n8n host and anomalous child process spawning (e.g., bash, sh, curl spawned by the n8n process).

What systems are affected by CVE-2026-25053?

This vulnerability affects the following AI/ML architecture patterns: agent frameworks, RAG pipelines, LLM-connected workflows, automation hubs, model serving, training pipelines.

What is the CVSS score for CVE-2026-25053?

CVE-2026-25053 has a CVSS v3.1 base score of 9.9 (CRITICAL). The EPSS exploitation probability is 0.57%.

What is the AI security impact?

Affected AI Architectures

agent frameworksRAG pipelinesLLM-connected workflowsautomation hubsmodel servingtraining pipelines

MITRE ATLAS Techniques

AML.T0012 Valid Accounts

AML.T0037 Data from Local System

AML.T0049 Exploit Public-Facing Application

AML.T0050 Command and Scripting Interpreter

AML.T0053 AI Agent Tool Invocation

AML.T0055 Unsecured Credentials

AML.T0072 Reverse Shell

AML.T0083 Credentials from AI Agent Configuration

AML.T0086 Exfiltration via AI Agent Tool Invocation

Compliance Controls Affected

EU AI Act: Art.15, Article 15

ISO 42001: A.6.2.6

NIST AI RMF: GOVERN-1.7, GOVERN-6.1, MANAGE-2.2

OWASP LLM Top 10: LLM02, LLM06, LLM07, LLM08

What are the technical details?

Original Advisory

n8n is an open source workflow automation platform. Prior to versions 1.123.10 and 2.5.0, vulnerabilities in the Git node allowed authenticated users with permission to create or modify workflows to execute arbitrary system commands or read arbitrary files on the n8n host. This issue has been patched in versions 1.123.10 and 2.5.0.

Exploitation Scenario

An attacker compromises or registers a low-privilege n8n account via credential stuffing, phishing, or abuse of self-registration. They create a workflow using the Git node with crafted parameters that inject OS commands—for example, reading /proc/self/environ to dump environment variables containing OPENAI_API_KEY, ANTHROPIC_API_KEY, database connection strings, and Stripe webhook secrets. With full command execution, they deploy a reverse shell, exfiltrate the entire n8n workflow database (containing all automation logic and embedded credentials), and use stolen AI API keys to pivot into LLM providers and downstream AI services. In AI-heavy deployments the attacker gains access to RAG database credentials, vector store API keys, and all integrated AI service configurations with zero forensic footprint in the AI provider logs.

Weaknesses (CWE)

CWE-78 Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection') Primary

CWE-78 — Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection'): The product constructs all or part of an OS command using externally-influenced input from an upstream component, but it does not neutralize or incorrectly neutralizes special elements that could modify the intended OS command when it is sent to a downstream component.

[Architecture and Design] If at all possible, use library calls rather than external processes to recreate the desired functionality.
[Architecture and Design, Operation] Run the code in a "jail" or similar sandbox environment that enforces strict boundaries between the process and the operating system. This may effectively restrict which files can be accessed in a particular directory or which commands can be executed by the software. OS-level examples include the Unix chroot jail, AppArmor, and SELinux. In general, managed code may provide some protection. For example, java.io.FilePermission in the Java SecurityManager allows the software to specify restrictions on file operations. This may not be a feasible solution, and it only limits the impact to the operating system; the rest of the application may still be subject to compromise. Be careful to avoid CWE-243 and other weaknesses related to jails.

Source: MITRE CWE corpus.