Jun 17, 2026·5 min read·22 visits
A validation flaw in the n8n Public API allowed users with read-only workflow permissions to execute retries, triggering unauthorized workflow executions on the server.
An incorrect authorization vulnerability in the Public API of n8n allows authenticated users with read-only permissions to bypass access control boundaries. By invoking the execution retry endpoint, an unauthorized user can trigger workflow executions, effectively escalating their privileges from workflow:read to workflow:execute.
The workflow automation platform n8n provides a Public API to manage and automate executions. In multi-tenant or team environments, n8n relies on a defined permission model to enforce boundaries between different users and projects. These boundaries ensure that users with read-only access cannot modify state or trigger actions.
A logical flaw in the Public API endpoint responsible for retrying workflow executions broke this authorization model. The endpoint allowed users to perform write/execution operations if they possessed read-only permissions. This flaw is classified under CWE-863 (Incorrect Authorization).
This vulnerability is significant for environments where workflows connect to production resources, cloud infrastructure, or databases. By abusing this endpoint, an unauthorized actor can trigger side effects in downstream integrations, leading to unintended modifications of third-party systems.
The root cause of this vulnerability lies in the incorrect mapping of permission scopes within the Public API router. n8n utilizes distinct scopes to govern user actions: workflow:read for viewing resources, workflow:write for configuration modifications, and workflow:execute for initiating runtime activities.
When a client targets the endpoint /v1/executions/{executionId}/retry, the application API gateway is responsible for verifying the caller's scope. Due to an integration oversight, the route guard applied to this endpoint verified the presence of the workflow:read scope instead of the workflow:execute scope.
Because the API key or authorization token of a read-only user naturally contains the workflow:read scope, the gateway passed the request to the execution controller. The backend then processed the execution payload as if the user possessed full execution rights, demonstrating a direct breakdown in role-based access control.
In the vulnerable version of the application, the controller responsible for execution actions implemented route protection decorators incorrectly. The following conceptual representation demonstrates how the middleware authorized the request based on the read scope:
// VULNERABLE ROUTE DEFINITION
@Post('/executions/:executionId/retry')
@Scope('workflow:read') // Bug: Validated read access instead of execute access
async retryExecution(@Param('executionId') executionId: string) {
return await this.executionService.retry(executionId);
}The fix applied by the n8n developers updated the decorator to reference the correct cryptographic execution scope. This change ensures that the router drops requests from read-only users before invoking the execution engine:
// PATCHED ROUTE DEFINITION
@Post('/executions/:executionId/retry')
@Scope('workflow:execute') // Fix: Enforce active execution permissions
async retryExecution(@Param('executionId') executionId: string) {
return await this.executionService.retry(executionId);
}This patch completely resolves the specific logical path. However, organizations should perform regression testing to ensure no other active state-changing endpoints are decorated with passive read scopes.
To exploit this vulnerability, an attacker requires a valid account or API key with read-only privileges. The attack vector is entirely network-based and can be executed via standard HTTP command-line tools.
First, the attacker identifies a target workflow to which they have read-only access. They query the execution list to discover historic execution identifiers:
curl -X GET "https://n8n.example.com/v1/executions?workflowId=10" \
-H "X-N8N-API-KEY: n8n_api_readonly_token_abc"Next, the attacker selects a target execution ID (e.g., 105) from the response. They issue a POST request to the retry endpoint using the same read-only credentials:
curl -X POST "https://n8n.example.com/v1/executions/105/retry" \
-H "X-N8N-API-KEY: n8n_api_readonly_token_abc" \
-H "Content-Type: application/json"The server returns a successful execution response, and the engine initiates a new runtime instance of the workflow using the historical input data of execution 105.
This vulnerability carries a CVSS v3.1 score of 6.4 (Medium). The scope parameter is classified as 'Changed' (S:C) because exploiting the API allows low-privileged users to trigger backend events that execute operations on third-party integrations outside the user's immediate scope.
Integrity impact is low to medium depending on the specific nodes configured in the target workflow. If the workflow executes SQL updates, writes to a cloud CRM, or interacts with transactional systems, repeating the execution can result in duplicated transactions or data corruption.
Resource exhaustion is also an operational risk. An automated script could repeatedly call the retry endpoint, causing n8n's execution queue to fill up. This would prevent legitimate workflows from running and deplete memory or CPU resources.
The definitive remediation for GHSA-H3JJ-5F3V-3685 is to upgrade n8n to a patched release. Systems on the 2.25.x release path must upgrade to 2.25.7 or later. Systems on the 2.26.x path must upgrade to 2.26.2 or later.
If patching cannot be performed immediately, administrators should implement strict network-level controls. Restrict access to the /v1/ API paths to trusted administrative source IP addresses using an API Gateway or Web Application Firewall.
Additionally, audit all active API keys and user privileges within the n8n application. Revoke any unnecessary keys and ensure that sharing workflows across teams is limited only to instances where collaboration is strictly required.
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:L/I:L/A:N| Product | Affected Versions | Fixed Version |
|---|---|---|
n8n n8n-io | < 2.25.7 | 2.25.7 |
n8n n8n-io | >= 2.26.0, < 2.26.2 | 2.26.2 |
| Attribute | Detail |
|---|---|
| CWE ID | CWE-863 |
| Attack Vector | Network |
| CVSS v3.1 Score | 6.4 (Medium) |
| CVSS Vector | CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:L/I:L/A:N |
| Exploit Status | poc |
| Patched Versions | 2.25.7, 2.26.2 |
The product performs an authorization check when an actor attempts to access a resource or perform an action, but it does not correctly associate the authorization with the explicit action being executed.
An incorrect authorization vulnerability (CWE-863) exists in the ASP.NET Core SignalR HTTP Connection Dispatcher (`Microsoft.AspNetCore.Http.Connections`). The connection negotiation process failed to bind the authenticated caller identity to the generated connection state, and subsequent transport or message dispatch handlers failed to verify caller consistency. Consequently, an authenticated low-privilege attacker who obtains a valid connection identifier or token can inject messages into or hijack another user's SignalR session, achieving an elevation of privilege across security boundaries.
CVE-2026-109478 is a network-accessible Denial of Service (DoS) vulnerability in ASP.NET Core caused by inefficient algorithmic complexity (CWE-407) when parsing comma-separated HTTP request headers. Prior to the patch, the internal helper method ParsingHelpers.GetHeaderSplit repeatedly called StringValues.Concat inside a loop iterating over each comma-separated token. Because StringValues allocates a new array of size k + 1 and copies all existing elements upon each concatenation, processing N tokens incurs O(N^2) time and space complexity. An unauthenticated remote attacker can send HTTP requests with headers containing thousands of short, comma-separated values to exhaust CPU cycles and induce severe Garbage Collection (GC) pauses, rendering the web application unresponsive.
CVE-2026-114147 is a high-severity Denial of Service (DoS) vulnerability affecting ASP.NET Core and specifically the Microsoft.AspNetCore.SignalR.StackExchangeRedis scale-out provider across .NET 8.0, 9.0, 10.0, and 11.0 preview releases. The vulnerability originates from a string comparison mismatch between internal per-connection group collections and the centralized Redis channel subscription registry, compounded by a race condition during connection teardown. Remote unauthenticated attackers can exploit this flaw to leave orphaned connection contexts and active Redis pub/sub listeners pinned in memory, leading to unbounded memory exhaustion and application process termination.
CVE-2026-114156 is an improper resolution of path equivalence vulnerability (CWE-41) in Microsoft ASP.NET Core affecting .NET 11.0 preview releases and Microsoft Visual Studio 2026. The vulnerability stems from a discrepancy between how path prefix validation helpers (such as PathString.StartsWithSegments) and downstream routing mechanisms handle alternate path separators and percent-encoded characters. An unauthenticated remote attacker can exploit this mismatch over the network to bypass path-based authorization middleware and access restricted endpoints.
CVE-2026-114157 is a high-severity local privilege escalation and security feature bypass vulnerability affecting ASP.NET Core on Microsoft Windows. The flaw resides in the Kestrel web server's Named Pipes transport layer (Microsoft.AspNetCore.Server.Kestrel.Transport.NamedPipes) and stems from a race condition (CWE-362) during broken pipe error recovery. When an active connection throws an IOException during handshake acceptance, the connection listener previously disposed of the broken NamedPipeServerStream instance before acquiring a replacement stream from its internal pool. This sequence created a momentary window where zero server instances of the named pipe existed in the Windows Named Pipe File System (NPFS). A local, low-privileged attacker can exploit this timing window to create a named pipe with the identical name and a permissive Discretionary Access Control List (DACL), hijacking subsequent inter-process communications and escalating privileges across security boundaries.
CVE-2026-124194 is a forced authentication and spoofing vulnerability in Microsoft Office Outlook caused by a missing user interface warning before dereferencing remote network paths (CWE-356). By embedding an attacker-controlled Universal Naming Convention (UNC) path within MAPI message properties such as PidLidReminderFileParameter, an unauthenticated remote attacker can force a victim's system to initiate an outbound SMB or WebDAV connection upon opening or previewing a crafted message. This automatic authentication handshake leaks the user's NetNTLMv2 hash, enabling offline password cracking or real-time NTLM relay attacks. The vulnerability was exploited in the wild as a zero-day prior to its remediation in Microsoft's November 2026 Patch Tuesday updates and is listed in the CISA Known Exploited Vulnerabilities (KEV) catalog.