CVE-2026-103255: n8n: Supabase node path traversal exposes admin APIs

CRITICAL
Published October 1, 2026
CISO Take

n8n's Supabase node inserts the tableId parameter into request paths without validation, so a workflow that binds tableId to untrusted input can be steered to the Auth and Storage APIs using the administrative serviceRole key, bypassing Row Level Security. Exploitation is rated high complexity because it needs a specific workflow design (attacker-controlled input mapped to tableId), and there is no public exploit, no Nuclei template, and no CISA KEV listing. EPSS is low at 0.003 (roughly the 80th percentile) and CISA SSVC says TRACK. The impact is still severe, because a successful hit gives full read and write access to user and storage data, and the scope change in the CVSS vector reflects that. Upgrade n8n to 1.123.80, 2.39.6 or 2.40.1 (depending on your branch), and in the meantime audit every workflow that feeds webhook, form or chat input into a Supabase node's tableId.

Sources:NVDEPSSGitHub Advisoryvulncheck.comATLASOpenSSF

What is the risk?

Rated critical (CVSS 9.0) with a changed scope, but real-world likelihood is moderate to low: attack complexity is high, no public exploit or scanner template exists, the CVE is not in KEV, and EPSS is 0.00296. Exposure depends on whether tableId is wired to untrusted input in an internet-reachable workflow, for example one triggered by a public webhook or an AI agent that picks tables dynamically. The serviceRole key makes the blast radius large, since it bypasses Row Level Security and reaches Auth and Storage. The n8n package has 244 other CVEs and an OpenSSF Scorecard of 6.7/10, which points to a recurring patching burden. Package risk score is 69/100.

How does the attack unfold?

Initial Access
Attacker finds an n8n workflow that exposes tableId to external input via a webhook, form or AI agent.
AML.T0049
Exploitation
A crafted tableId with path traversal sequences redirects the Supabase node request to the Auth or Storage API.
AML.T0053
Privilege Abuse
The request is sent with the administrative serviceRole key, bypassing Row Level Security.
AML.T0055
Impact
Attacker reads or modifies user accounts and stored files, exposing sensitive data and corrupting RAG sources.
AML.T0036

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
9.0 / 10
EPSS
0.3%
chance of exploitation in 30 days
Higher than 23% of all CVEs
Exploitation Status
No known exploitation
Sophistication
Moderate

What is the attack surface?

AVACPRUISCIA
AVNetwork
ACHigh
PRNone
UINone
SChanged
CHigh
IHigh
AHigh

What should I do?

1 step
  1. 1) Upgrade n8n to 1.123.80 (1.x branch), 2.39.6 (2.0.0 to 2.39.x), or 2.40.1 (2.40.0). 2) Until patched, review all workflows with Supabase nodes and hard-code tableId or validate it against an allowlist, never binding it to webhook, form, chat or LLM output. 3) Replace the serviceRole key with a scoped, RLS-enforced role wherever the workflow does not need admin access. 4) Rotate the Supabase serviceRole key if any exposed workflow could have been hit. 5) Detection: review Supabase API logs for requests to /auth/v1 and /storage/v1 coming from the n8n host's IP, and look for path-traversal sequences such as ../ or encoded variants in n8n execution inputs. 6) Restrict public webhook exposure and put authentication in front of workflows that accept external input.

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?

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act
Art. 15-Accuracy, robustness and cybersecurity
ISO 42001
A.7.4-Quality of data for AI systems
NIST AI RMF
MANAGE 2.3-Procedures to respond to and recover from unknown risks
OWASP LLM Top 10
LLM06-Excessive Agency

Frequently Asked Questions

What is CVE-2026-103255?

n8n's Supabase node inserts the tableId parameter into request paths without validation, so a workflow that binds tableId to untrusted input can be steered to the Auth and Storage APIs using the administrative serviceRole key, bypassing Row Level Security. Exploitation is rated high complexity because it needs a specific workflow design (attacker-controlled input mapped to tableId), and there is no public exploit, no Nuclei template, and no CISA KEV listing. EPSS is low at 0.003 (roughly the 80th percentile) and CISA SSVC says TRACK. The impact is still severe, because a successful hit gives full read and write access to user and storage data, and the scope change in the CVSS vector reflects that. Upgrade n8n to 1.123.80, 2.39.6 or 2.40.1 (depending on your branch), and in the meantime audit every workflow that feeds webhook, form or chat input into a Supabase node's tableId.

Is CVE-2026-103255 actively exploited?

No confirmed active exploitation of CVE-2026-103255 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-103255?

1) Upgrade n8n to 1.123.80 (1.x branch), 2.39.6 (2.0.0 to 2.39.x), or 2.40.1 (2.40.0). 2) Until patched, review all workflows with Supabase nodes and hard-code tableId or validate it against an allowlist, never binding it to webhook, form, chat or LLM output. 3) Replace the serviceRole key with a scoped, RLS-enforced role wherever the workflow does not need admin access. 4) Rotate the Supabase serviceRole key if any exposed workflow could have been hit. 5) Detection: review Supabase API logs for requests to /auth/v1 and /storage/v1 coming from the n8n host's IP, and look for path-traversal sequences such as ../ or encoded variants in n8n execution inputs. 6) Restrict public webhook exposure and put authentication in front of workflows that accept external input.

What systems are affected by CVE-2026-103255?

This vulnerability affects the following AI/ML architecture patterns: Agent frameworks, Workflow automation platforms, RAG pipelines, Vector database integrations, LLM-driven automation backends.

What is the CVSS score for CVE-2026-103255?

CVE-2026-103255 has a CVSS v3.1 base score of 9.0 (CRITICAL). The EPSS exploitation probability is 0.30%.

What is the AI security impact?

Affected AI Architectures

Agent frameworksWorkflow automation platformsRAG pipelinesVector database integrationsLLM-driven automation backends

MITRE ATLAS Techniques

AML.T0036Data from Information Repositories
AML.T0049Exploit Public-Facing Application
AML.T0053AI Agent Tool Invocation
AML.T0055Unsecured Credentials

Compliance Controls Affected

EU AI Act: Art. 15
ISO 42001: A.7.4
NIST AI RMF: MANAGE 2.3
OWASP LLM Top 10: LLM06

What are the technical details?

Original Advisory

n8n versions before 1.123.80, from 2.0.0 before 2.39.6, and from 2.40.0 before 2.40.1 contain a path traversal vulnerability in the Supabase node where the tableId parameter is inserted into request paths without validation. Attackers can exploit workflows binding tableId to untrusted input to traverse to Auth and Storage APIs using the administrative serviceRole key, bypassing Row Level Security and enabling unauthorized data access and modification.

Exploitation Scenario

An organization runs an n8n workflow behind a public webhook that lets users query a Supabase-backed knowledge base, with the table name taken from the request or chosen by an AI agent. The workflow uses the Supabase serviceRole credential. An attacker sends a crafted tableId containing path traversal sequences, which redirects the request from the REST table endpoint to the Auth admin or Storage API. Because the request carries the admin key, it returns the user list or bucket contents, and can modify or delete users and files. The attacker then uses that data for account takeover or to read stored documents, including RAG source files.

Weaknesses (CWE)

CWE-73 — External Control of File Name or Path: The product allows user input to control or influence paths or file names that are used in filesystem operations.

  • [Architecture and Design] When the set of filenames is limited or known, create a mapping from a set of fixed input values (such as numeric IDs) to the actual filenames, and reject all other inputs. For example, ID 1 could map to "inbox.txt" and ID 2 could map to "profile.txt". Features such as the ESAPI AccessReferenceMap provide this capability.
  • [Architecture and Design, Operation] Run your code in a "jail" or similar sandbox environment that enforces strict boundaries between the process and the operating system. This may effectively restrict all access to files within a particular directory. Examples include the Unix chroot jail and AppArmor. In general, managed code may provide some protection. This may not be a feasible solution, and it only limits the impact to the operating system; the rest of your application may still be subject to compromise. Be careful to avoid CWE-243 and other weaknesses related to jails.

Source: MITRE CWE corpus.

CVSS Vector

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

Timeline

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

Related Vulnerabilities