CVE-2026-77072: n8n stored XSS — HIGH

CISO Take

A stored cross-site scripting flaw in n8n's Form node lets an authenticated workspace member bypass the completion page's sandboxing CSP by setting respondWith to 'redirect' via an expression while leaving responseText populated, causing the page to render raw unsanitized HTML and script from the n8n origin. This matters because n8n is a widely deployed AI agent/workflow automation platform (16 tracked downstream dependents, 201 other CVEs in the package, OpenSSF Scorecard 6.6/10), and public-facing Form nodes are commonly used as intake front doors for AI agent workflows, meaning any external visitor who submits the resulting form gets attacker script executed same-origin with their own session. There is no CISA KEV listing, no published EPSS score, and no public exploit or Nuclei template observed, so this is not currently under mass exploitation — but the exploitation bar is low (a single malicious or compromised member account, no special tooling) and the blast radius extends to any third party who interacts with an exposed public form. Upgrade immediately to n8n 1.123.69, 2.33.4, or 2.34.1 (or later), and in the meantime audit existing workflows for Form nodes using respondWith='redirect' with populated responseText fields, and tighten who can edit public-facing workflows.

Sources:NVDGitHub Advisoryvulncheck.comOpenSSFATLAS

What is the risk?

Rated high severity by the vendor though no CVSS vector or EPSS score is published. Exploitability is moderate: it requires an authenticated n8n workspace member (not an anonymous internet attacker) to deliberately or accidentally misconfigure a Form node, which narrows the threat actor pool to insiders, compromised member accounts, or supply-chain-tainted workflow templates. Once configured, however, exploitation of end users is trivial and automatic — any visitor submitting the public form is affected with no further action needed. No KEV listing, EPSS score, public PoC, or Nuclei template exists today, indicating exploitation is not currently observed in the wild, but the low technical bar (a single expression change) and the fact that victims are external, uncontrolled third parties (not the n8n operator) elevate real-world risk beyond what the absence of a CVSS score suggests.

How does the attack unfold?

Malicious configuration

An authenticated n8n workspace member sets a public Form node's respondWith to 'redirect' via expression while keeping responseText populated with an XSS payload.

AML.T0081

CSP bypass persists

Because sandboxing CSP is only applied when respondWith is not 'redirect', the malicious configuration silently persists in the workflow undetected.

Victim interaction

Any external visitor who submits the public-facing form triggers the completion page to render the unsanitized HTML/script from the n8n origin.

AML.T0078

Session compromise

The injected script executes in the visitor's browser same-origin with their own n8n session, enabling session token theft or phishing.

Malicious configuration

An authenticated n8n workspace member sets a public Form node's respondWith to 'redirect' via expression while keeping responseText populated with an XSS payload.

AML.T0081

CSP bypass persists

Because sandboxing CSP is only applied when respondWith is not 'redirect', the malicious configuration silently persists in the workflow undetected.

Victim interaction

Any external visitor who submits the public-facing form triggers the completion page to render the unsanitized HTML/script from the n8n origin.

AML.T0078

Session compromise

The injected script executes in the visitor's browser same-origin with their own n8n session, enabling session token theft or phishing.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
n8n	npm	—	No patch
209.2KOpenSSF 7.116 dependentsPushed today44% patched~19d to patchFull package profile →

Do you use n8n? You're affected.

How severe is it?

CVSS 3.1

N/A

EPSS

0.3%

chance of exploitation in 30 days

Higher than 23% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

No known exploitation

Sophistication

Moderate

What should I do?

1 step

Patch n8n to 1.123.69, 2.33.4, or 2.34.1 (or later) immediately — this is the only complete fix since the vulnerability is in core completion-page rendering logic. Until patched: audit all workflows for Form nodes where respondWith is set (especially via expression) to 'redirect' while responseText is populated, and disable or unpublish affected public forms. Reduce the pool of workspace members who can edit public-facing workflows (least privilege / role review), since exploitation requires an authenticated member. For detection, review workflow change history/audit logs for edits to Form node respondWith/responseText fields, and monitor completion-page HTTP responses for unexpected inline script content. After patching, re-verify that the completion-page CSP is enforced across all respondWith modes, not just the default.

What does CISA's SSVC say?

DecisionTrack

Exploitationnone

AutomatableNo

Technical Impacttotal

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

How is it classified?

Code ExecutionPrivacy ViolationAgentPluginAML.T0078- Drive-by Compromise AML.T0081- Modify AI Agent Configuration

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Article 15-Accuracy, robustness and cybersecurity

ISO 42001

A.6.2.6-AI system security and vulnerability management

NIST AI RMF

GOVERN 6.1-Third-party AI system risk management

OWASP LLM Top 10

LLM02-Insecure Output Handling

Frequently Asked Questions

What is CVE-2026-77072?

A stored cross-site scripting flaw in n8n's Form node lets an authenticated workspace member bypass the completion page's sandboxing CSP by setting respondWith to 'redirect' via an expression while leaving responseText populated, causing the page to render raw unsanitized HTML and script from the n8n origin. This matters because n8n is a widely deployed AI agent/workflow automation platform (16 tracked downstream dependents, 201 other CVEs in the package, OpenSSF Scorecard 6.6/10), and public-facing Form nodes are commonly used as intake front doors for AI agent workflows, meaning any external visitor who submits the resulting form gets attacker script executed same-origin with their own session. There is no CISA KEV listing, no published EPSS score, and no public exploit or Nuclei template observed, so this is not currently under mass exploitation — but the exploitation bar is low (a single malicious or compromised member account, no special tooling) and the blast radius extends to any third party who interacts with an exposed public form. Upgrade immediately to n8n 1.123.69, 2.33.4, or 2.34.1 (or later), and in the meantime audit existing workflows for Form nodes using respondWith='redirect' with populated responseText fields, and tighten who can edit public-facing workflows.

Is CVE-2026-77072 actively exploited?

No confirmed active exploitation of CVE-2026-77072 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-77072?

Patch n8n to 1.123.69, 2.33.4, or 2.34.1 (or later) immediately — this is the only complete fix since the vulnerability is in core completion-page rendering logic. Until patched: audit all workflows for Form nodes where respondWith is set (especially via expression) to 'redirect' while responseText is populated, and disable or unpublish affected public forms. Reduce the pool of workspace members who can edit public-facing workflows (least privilege / role review), since exploitation requires an authenticated member. For detection, review workflow change history/audit logs for edits to Form node respondWith/responseText fields, and monitor completion-page HTTP responses for unexpected inline script content. After patching, re-verify that the completion-page CSP is enforced across all respondWith modes, not just the default.

What systems are affected by CVE-2026-77072?

This vulnerability affects the following AI/ML architecture patterns: agent frameworks, public-facing intake forms feeding agent workflows.

What is the CVSS score for CVE-2026-77072?

No CVSS score has been assigned yet.

What is the AI security impact?

Affected AI Architectures

agent frameworkspublic-facing intake forms feeding agent workflows

MITRE ATLAS Techniques

AML.T0078Drive-by Compromise

AML.T0081Modify AI Agent Configuration

Compliance Controls Affected

EU AI Act: Article 15

ISO 42001: A.6.2.6

NIST AI RMF: GOVERN 6.1

OWASP LLM Top 10: LLM02

What are the technical details?

Original Advisory

n8n before 1.123.69, 2.33.4, and 2.34.1 contains a stored cross-site scripting vulnerability in the Form node's completion page. The completion page applied its sandboxing Content-Security-Policy only when respondWith was not set to 'redirect', but responseText was always rendered as raw HTML. An authenticated member could set respondWith to 'redirect' via an expression while keeping responseText populated, causing the completion page to serve unsanitized HTML and script from the n8n origin. Any visitor who submitted the resulting public form would have that script execute same-origin with their session.

Exploitation Scenario

A malicious or compromised authenticated member of an n8n workspace edits a public-facing Form node workflow (e.g., a customer intake form that feeds an AI agent pipeline). They set respondWith to 'redirect' using an n8n expression while leaving the responseText field populated with an XSS payload disguised as normal completion text. Because the sandboxing CSP is only applied when respondWith is not 'redirect', the completion page silently serves the payload as raw, unsanitized HTML from the trusted n8n origin. When any external user submits that public form — a customer, partner, or job applicant with no reason to suspect anything — the injected script executes in their browser same-origin with their own session, enabling session token theft, credential-phishing overlays, or redirection to attacker infrastructure, all while appearing to come from the legitimate n8n-hosted form.

Weaknesses (CWE)

CWE-79Improper Neutralization of Input During Web Page GenerationPrimary CWE-79Improper Neutralization of Input During Web Page Generation

CWE-79 — Improper Neutralization of Input During Web Page Generation: The product does not neutralize or incorrectly neutralizes user-controllable input before it is placed in output that is used as a web page that is served to other users.

[Architecture and Design] Use a vetted library or framework that does not allow this weakness to occur or provides constructs that make this weakness easier to avoid [REF-1482]. Examples of libraries and frameworks that make it easier to generate properly encoded output include Microsoft's Anti-XSS library, the OWASP ESAPI Encoding module, and Apache Wicket.
[Implementation, Architecture and Design] Understand the context in which your data will be used and the encoding that will be expected. This is especially important when transmitting data between different components, or when generating outputs that can contain multiple encodings at the same time, such as web pages or multi-part mail messages. Study all expected communication protocols and data representations to determine the required encoding strategies. For any data that will be output to another web page, especially any data that was received from external inputs, use the appropriate encoding on all non-alphanumeric characters. Parts of the same output document may require different encodings, which will vary depending on whether the output is in the: etc. Note that HTML Entity Encoding is only appropriate for the HTML body. Consult the XSS Prevention Cheat Sheet [REF-724] for more details on the types of encoding and escaping that are needed. HTML body Element attributes (such as src="XYZ") URIs JavaScript sections Casca

Source: MITRE CWE corpus.