CVE-2026-103248: n8n: Supabase node filter injection exposes tables
CRITICALn8n's Supabase node fails to escape field values in its Filters (String) mode, so untrusted input can be turned into PostgREST filter expressions that read, update, or delete every row in a table with a single request. The CVSS score is 9.0 (critical) with a changed scope, because the injected filters run with the Supabase credentials stored in n8n rather than the caller's own privileges. Near-term exploitation risk is moderate: EPSS is 0.003 (79th percentile), the CVE is not in CISA KEV, no public exploit or Nuclei template exists, and CISA SSVC says TRACK. The high attack complexity reflects that only workflows passing attacker-controlled data into that filter mode are exposed. n8n has 244 other CVEs on record and a Package Risk Score of 69/100, so treat this as part of a recurring pattern. Upgrade to 1.123.80, 2.39.6, or 2.40.1 (depending on your branch) and audit workflows that use the Supabase node with webhook, form, or chat inputs.
What is the risk?
Critical severity (CVSS 9.0), but real-world risk depends on workflow design. Exploitation needs no privileges or user interaction and is network-reachable, but it requires a workflow that feeds untrusted input into Supabase Filters (String) mode, hence high complexity. Exploitation likelihood is currently low to moderate (EPSS 0.003, 79th percentile, not in KEV, no public PoC, SSVC TRACK). Impact is severe where it applies: full confidentiality, integrity, and availability loss on the targeted Supabase tables, because n8n typically uses a service-role key that bypasses row-level security. The risk rises quickly if a public proof of concept appears, since injection techniques against PostgREST filters are well understood.
How does the attack unfold?
What systems are affected?
| Package | Ecosystem | Vulnerable Range | Patched |
|---|---|---|---|
| n8n | npm | — | No patch |
Do you use n8n? You're affected.
How severe is it?
What is the attack surface?
What should I do?
1 step1) Patch to n8n 1.123.80, 2.39.6, or 2.40.1 or later, depending on your release branch. 2) Inventory all workflows that use the Supabase node and identify those using Filters (String) mode with values derived from external or LLM-generated input. 3) Until patched, switch to the structured filter UI or validate and allowlist input before it reaches the node (reject characters such as commas, parentheses, and periods that carry PostgREST filter syntax). 4) Use least-privilege Supabase credentials: avoid the service-role key, apply row-level security, and scope the key to only the required tables and operations. 5) Take Supabase backups and enable point-in-time recovery so mass update or delete is recoverable. 6) Detect abuse by reviewing Supabase/PostgREST logs for unusually broad queries, DELETE or PATCH requests without narrow filters, and n8n execution logs for suspicious input in webhook-triggered workflows. 7) Restrict public exposure of n8n webhooks and forms where possible.
What does CISA's SSVC say?
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:
Frequently Asked Questions
What is CVE-2026-103248?
n8n's Supabase node fails to escape field values in its Filters (String) mode, so untrusted input can be turned into PostgREST filter expressions that read, update, or delete every row in a table with a single request. The CVSS score is 9.0 (critical) with a changed scope, because the injected filters run with the Supabase credentials stored in n8n rather than the caller's own privileges. Near-term exploitation risk is moderate: EPSS is 0.003 (79th percentile), the CVE is not in CISA KEV, no public exploit or Nuclei template exists, and CISA SSVC says TRACK. The high attack complexity reflects that only workflows passing attacker-controlled data into that filter mode are exposed. n8n has 244 other CVEs on record and a Package Risk Score of 69/100, so treat this as part of a recurring pattern. Upgrade to 1.123.80, 2.39.6, or 2.40.1 (depending on your branch) and audit workflows that use the Supabase node with webhook, form, or chat inputs.
Is CVE-2026-103248 actively exploited?
No confirmed active exploitation of CVE-2026-103248 has been reported, but organizations should still patch proactively.
How to fix CVE-2026-103248?
1) Patch to n8n 1.123.80, 2.39.6, or 2.40.1 or later, depending on your release branch. 2) Inventory all workflows that use the Supabase node and identify those using Filters (String) mode with values derived from external or LLM-generated input. 3) Until patched, switch to the structured filter UI or validate and allowlist input before it reaches the node (reject characters such as commas, parentheses, and periods that carry PostgREST filter syntax). 4) Use least-privilege Supabase credentials: avoid the service-role key, apply row-level security, and scope the key to only the required tables and operations. 5) Take Supabase backups and enable point-in-time recovery so mass update or delete is recoverable. 6) Detect abuse by reviewing Supabase/PostgREST logs for unusually broad queries, DELETE or PATCH requests without narrow filters, and n8n execution logs for suspicious input in webhook-triggered workflows. 7) Restrict public exposure of n8n webhooks and forms where possible.
What systems are affected by CVE-2026-103248?
This vulnerability affects the following AI/ML architecture patterns: Agent frameworks, Workflow automation platforms, RAG pipelines, Vector databases, Chatbot backends.
What is the CVSS score for CVE-2026-103248?
CVE-2026-103248 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
MITRE ATLAS Techniques
AML.T0049Exploit Public-Facing ApplicationAML.T0053AI Agent Tool InvocationAML.T0085.001AI Agent ToolsAML.T0101Data Destruction via AI Agent Tool InvocationCompliance Controls Affected
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 filter injection vulnerability in the Supabase node's Filters (String) mode that fails to escape field values. Attackers can inject filter expressions from untrusted input to read all table rows, update all records, or delete entire tables in a single request.
Exploitation Scenario
An organization runs an n8n AI support agent that receives chat messages via a public webhook, extracts a customer identifier, and looks up the customer's records in Supabase using a Filters (String) expression. The attacker submits an identifier containing PostgREST filter syntax, which is not escaped, so the condition is rewritten to match all rows. A single request returns the entire customers or conversation table. Using a workflow that updates or deletes records, the attacker modifies the same pattern to overwrite or delete all rows, poisoning the agent's knowledge base or wiping the table. Because n8n connects with its own credentials, row-level security is typically bypassed.
Weaknesses (CWE)
CWE-89Improper Neutralization of Special Elements used in an SQL CommandPrimaryCWE-89Improper Neutralization of Special Elements used in an SQL CommandCWE-89 — Improper Neutralization of Special Elements used in an SQL Command: The product constructs all or part of an SQL command using externally-influenced input from an upstream component, but it does not neutralize or incorrectly neutralizes special elements that could modify the intended SQL command when it is sent to a downstream component. Without sufficient removal or quoting of SQL syntax in user-controllable inputs, the generated SQL query can cause those inputs to be interpreted as SQL instead of ordinary user data.
- [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]. For example, consider using persistence layers such as Hibernate or Enterprise Java Beans, which can provide significant protection against SQL injection if used properly.
- [Architecture and Design] If available, use structured mechanisms that automatically enforce the separation between data and code. These mechanisms may be able to provide the relevant quoting, encoding, and validation automatically, instead of relying on the developer to provide this capability at every point where output is generated. Process SQL queries using prepared statements, parameterized queries, or stored procedures. These features should accept parameters or variables and support strong typing. Do not dynamically construct and execute query strings within these features using "exec" or similar functionality, since this may re-introduce the possibility of SQL injection. [REF-867]
Source: MITRE CWE corpus.
CVSS Vector
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:C/C:H/I:H/A:HReferences
Timeline
Related Vulnerabilities
CVE-2026-3366310.0n8n: member role steals plaintext HTTP credentials
Same package: n8nCVE-2026-3366010.0TensorFlow: type confusion NPD in tensor conversion
Same package: n8nCVE-2026-2185810.0n8n: Input Validation flaw enables exploitation
Same package: n8nCVE-2026-275779.9n8n: Code Injection enables RCE
Same package: n8nCVE-2026-274949.9n8n: security flaw enables exploitation
Same package: n8n