CVE-2026-103260: n8n: approval bypass in Send and Wait chat mode

MEDIUM
Published October 1, 2026
CISO Take

n8n's Send and Wait node in Approve Within Chat mode does not verify who submits a resume request, so an unauthenticated party can push a paused workflow past its human approval gate and trigger the guarded actions that follow. Technical severity is limited (CVSS 4.0, high attack complexity, integrity impact only), and exploitation signals are weak: EPSS 0.00281 (top 81st percentile), no public exploit, no Nuclei template, not in CISA KEV, and CISA SSVC says TRACK. The practical risk is higher than the score suggests for organizations that use this node as a control for AI agent actions such as payments, deployments, data deletion or outbound messages, because the approval is the only human safeguard. Upgrade to n8n 2.39.6 or 2.40.1 or later, and review executions that resumed without a matching approver identity. Treat this as a scheduled patch rather than a breaking-alert item, but prioritize workflows where an agent's high-impact tool calls depend on this approval step.

Sources:NVDEPSSGitHub AdvisoryOpenSSFATLASvulncheck.com

What is the risk?

Overall risk is LOW to MEDIUM. The CVSS vector (AV:N/AC:H/PR:N/UI:N/S:C/C:N/I:L/A:N) shows network reach with no authentication but high complexity, probably because an attacker must know or guess the waiting execution's resume endpoint or token. EPSS is low at 0.00281, there is no public exploit or scanner template, and the CVE is not in KEV. The scope-changed rating reflects that the bypass affects downstream actions beyond the node. Real-world impact depends entirely on what the guarded actions do. The package has 244 other CVEs, which signals a heavy ongoing vulnerability load, and a 6.7/10 OpenSSF Scorecard shows moderate maturity. The weakness is CWE-862 (Missing Authorization).

How does the attack unfold?

Reconnaissance
Attacker identifies an exposed n8n instance and a waiting Send and Wait execution awaiting chat approval.
AML.T0006
Initial Access
Attacker sends an unauthenticated resume request to the vulnerable endpoint on an unpatched n8n version.
AML.T0049
Approval Bypass
The node accepts the request without verifying requester identity or approval permission and advances the execution.
AML.T0053
Impact
Guarded downstream actions run without legitimate human approval, undermining integrity of agent-driven operations.

What systems are affected?

PackageEcosystemVulnerable RangePatched
n8nnpm—No patch
210.5KOpenSSF 7.116 dependentsPushed 3d ago42% patched~19d to patchFull package profile →

Do you use n8n? You're affected.

How severe is it?

CVSS 3.1
4.0 / 10
EPSS
0.3%
chance of exploitation in 30 days
Higher than 21% of all CVEs
Exploitation Status
No known exploitation
Sophistication
Moderate

What is the attack surface?

AVACPRUISCIA
AVNetwork
ACHigh
PRNone
UINone
SChanged
CNone
ILow
ANone

What should I do?

1 step
  1. 1) Upgrade n8n to 2.39.6 or later on the 2.39 line, or 2.40.1 or later on the 2.40 line. 2) Until patched, do not rely on Approve Within Chat as the only gate for high-impact actions. Add a second control such as an authenticated approval through a different channel, or a separate authorization check inside the workflow before the sensitive step. 3) Restrict network exposure of the n8n webhook and resume endpoints with a reverse proxy, IP allowlisting or VPN, and avoid exposing waiting-execution URLs in public channels. 4) Detection: audit execution logs for Send and Wait executions that resumed without a corresponding approver identity or at unexpected times or source IPs, and look for resume requests from unfamiliar addresses. 5) Review GHSA-342v-gmh6-738j and the VulnCheck advisory for any additional guidance, and apply least privilege to the credentials and tools the guarded actions use.

What does CISA's SSVC say?

DecisionTrack
Exploitationnone
AutomatableNo
Technical Impactpartial

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

How is it classified?

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act
Article 14-Human Oversight
ISO 42001
A.9.2-Processes for responsible use of AI systems
NIST AI RMF
GOVERN 1.4-Risk management process and controls
OWASP LLM Top 10
LLM06-Excessive Agency

Frequently Asked Questions

What is CVE-2026-103260?

n8n's Send and Wait node in Approve Within Chat mode does not verify who submits a resume request, so an unauthenticated party can push a paused workflow past its human approval gate and trigger the guarded actions that follow. Technical severity is limited (CVSS 4.0, high attack complexity, integrity impact only), and exploitation signals are weak: EPSS 0.00281 (top 81st percentile), no public exploit, no Nuclei template, not in CISA KEV, and CISA SSVC says TRACK. The practical risk is higher than the score suggests for organizations that use this node as a control for AI agent actions such as payments, deployments, data deletion or outbound messages, because the approval is the only human safeguard. Upgrade to n8n 2.39.6 or 2.40.1 or later, and review executions that resumed without a matching approver identity. Treat this as a scheduled patch rather than a breaking-alert item, but prioritize workflows where an agent's high-impact tool calls depend on this approval step.

Is CVE-2026-103260 actively exploited?

No confirmed active exploitation of CVE-2026-103260 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-103260?

1) Upgrade n8n to 2.39.6 or later on the 2.39 line, or 2.40.1 or later on the 2.40 line. 2) Until patched, do not rely on Approve Within Chat as the only gate for high-impact actions. Add a second control such as an authenticated approval through a different channel, or a separate authorization check inside the workflow before the sensitive step. 3) Restrict network exposure of the n8n webhook and resume endpoints with a reverse proxy, IP allowlisting or VPN, and avoid exposing waiting-execution URLs in public channels. 4) Detection: audit execution logs for Send and Wait executions that resumed without a corresponding approver identity or at unexpected times or source IPs, and look for resume requests from unfamiliar addresses. 5) Review GHSA-342v-gmh6-738j and the VulnCheck advisory for any additional guidance, and apply least privilege to the credentials and tools the guarded actions use.

What systems are affected by CVE-2026-103260?

This vulnerability affects the following AI/ML architecture patterns: Agent frameworks, Human-in-the-loop approval workflows, Workflow automation platforms, LLM tool-calling pipelines.

What is the CVSS score for CVE-2026-103260?

CVE-2026-103260 has a CVSS v3.1 base score of 4.0 (MEDIUM). The EPSS exploitation probability is 0.28%.

What is the AI security impact?

Affected AI Architectures

Agent frameworksHuman-in-the-loop approval workflowsWorkflow automation platformsLLM tool-calling pipelines

MITRE ATLAS Techniques

AML.T0049Exploit Public-Facing Application
AML.T0053AI Agent Tool Invocation

Compliance Controls Affected

EU AI Act: Article 14
ISO 42001: A.9.2
NIST AI RMF: GOVERN 1.4
OWASP LLM Top 10: LLM06

What are the technical details?

Original Advisory

n8n versions before 2.39.6 and 2.40.0 before 2.40.1 contain an approval bypass vulnerability in the Send and Wait node's Approve Within Chat mode. Attackers can submit resume requests without verification of the requester's identity or approval permissions, allowing unauthenticated users to advance waiting executions and trigger guarded actions.

Exploitation Scenario

An organization runs an n8n AI agent workflow that drafts a vendor payment or production change and then pauses at a Send and Wait node in Approve Within Chat mode, waiting for a manager's approval. An external attacker who learns or infers the resume endpoint for a waiting execution (for example from a leaked link, logs or a shared chat transcript) sends a crafted resume request. Because the node does not verify the requester's identity or approval permission, the execution advances as if approved. The agent then carries out the guarded action, such as issuing a payment instruction, modifying a record or sending an external message, without any legitimate human sign-off.

Weaknesses (CWE)

CWE-862 — Missing Authorization: The product does not perform an authorization check when an actor attempts to access a resource or perform an action.

  • [Architecture and Design] Divide the product into anonymous, normal, privileged, and administrative areas. Reduce the attack surface by carefully mapping roles with data and functionality. Use role-based access control (RBAC) [REF-229] to enforce the roles at the appropriate boundaries. Note that this approach may not protect against horizontal authorization, i.e., it will not protect a user from attacking others with the same role.
  • [Architecture and Design] Ensure that access control checks are performed related to the business logic. These checks may be different than the access control checks that are applied to more generic resources such as files, connections, processes, memory, and database records. For example, a database may restrict access for medical records to a specific database user, but each record might only be intended to be accessible to the patient and the patient's doctor [REF-7].

Source: MITRE CWE corpus.

CVSS Vector

CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:C/C:N/I:L/A:N

Timeline

Published
October 1, 2026
Last Modified
October 1, 2026
First Seen
October 1, 2026

Related Vulnerabilities