/

CVE-2026-42861

Severity:

CRITICAL RISK

(9.6)

Mass Assignment (CWE-915) in FlowiseAI Variables API

CVE-2026-42861

Analyzed

Vulnerability type

Category:

Mass Assignment / Improper Authorization

CWE:

CWE-284, CWE-639, CWE-915

Impact:

Cross-Tenant Access and Isolation Bypass

Severity:

CRITICAL RISK

(9.6)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

FlowiseAI's variable update endpoint suffers from a mass assignment flaw that lets authenticated users modify server-controlled fields like workspaceId, createdDate, and updatedDate.

In multi-workspace deployments, this can allow a user to reassign variables into arbitrary workspaces, potentially violating tenant isolation and exposing cross-tenant data or configuration.

Affected context

This vulnerability appears in FlowiseAI deployments where:

  • Multiple workspaces or tenants share the same backend
  • The variable update endpoint accepts JSON bodies that include server-managed fields
  • The backend lacks strict server-side validation and ownership checks on workspaceId and other metadata

It is most impactful in environments relying on workspace boundaries for data separation and access control.

Exploit preconditions

  • Attack surface: HTTP API endpoint for updating variable resources
  • Required access: Any authenticated account capable of updating variables
  • Trigger point: Sending an update request that includes modified server-controlled fields such as workspaceId
  • Why exploitation is possible: The backend trusts client-supplied fields without enforcing that workspaceId or related properties belong to the caller's authorized workspace, and performs a broad object update (mass assignment) instead of whitelisting allowed attributes.

Why this happens

The core issue is a mass assignment pattern where the server maps client-provided JSON directly onto a persistent model, including fields intended to be server-controlled.

Without explicit whitelisting or per-field validation, properties such as workspaceId, createdDate, and updatedDate are overwritten based on user input.

This maps to CWE-915: Improperly Controlled Modification of Dynamically-Determined Object Attributes, and is compounded by missing or weak authorization checks (CWE-284 / CWE-639) on workspace ownership and resource scoping.

Potential impact

  • Users could reassign variables into workspaces they do not own or control
  • Cross-tenant data or configuration exposure in multi-workspace environments
  • Breakage of tenant isolation guarantees and potential regulatory compliance issues
  • Confusion or corruption of workspace-level configurations and flows
  • Indirect privilege escalation via access to variables used by higher-privileged flows

General mitigation guidance

Resilience against this class of issue relies on strict control of which fields clients may modify and enforcing authorization at the resource and tenant boundary.

Common defensive patterns include:

  • Implementing explicit allowlists of updatable fields for each endpoint, excluding server-controlled metadata such as workspaceId, createdDate, and updatedDate
  • Performing server-side ownership and tenant checks on every variable update, ensuring the resource and its workspace belong to the authenticated user or role
  • Separating client-visible DTOs from persistence models so internal fields are never directly bound from request bodies
  • Centralizing authorization logic to consistently enforce workspace and tenant isolation rules
  • Applying least-privilege principles to API operations, ensuring only necessary attributes are mutable through public endpoints

How teams detect issues like this

Organizations can detect similar vulnerabilities by:

  • Using static analysis or linters to flag patterns where request bodies are directly bound to ORM entities or models without field whitelisting
  • Reviewing controllers and handlers for update operations that accept entire objects instead of specific, validated attributes
  • Implementing security review checklists that explicitly examine multi-tenant boundaries, including workspaceId or tenantId handling
  • Adding runtime logging and anomaly detection to spot suspicious cross-workspace updates or sudden changes in workspace ownership of variables
  • Employing automated API security testing to probe for insecure direct object references and mass assignment on multi-tenant resources

Sources

SourceLink
github.comgithub.com/FlowiseAI/Flowise/releases/tag/flowise%403.1.2
github.comgithub.com/FlowiseAI/Flowise/security/advisories/GHSA-6fw7-3q8r-m5vj

Frequently asked

questions

Yes, it is exploitable remotely through the FlowiseAI HTTP API by any authenticated user who can reach the variable update endpoint.

The attacker needs valid authentication with permissions to update variables in at least one workspace; no elevated administrative privileges are required to attempt cross-workspace reassignment if checks are missing.

Implement strict field-level whitelisting for update endpoints, ensure server-controlled fields (such as workspaceId and timestamps) are never taken from client input, and enforce robust authorization checks that validate ownership and tenant boundaries on every update.

Adopt a defensive design that separates external request models from internal persistence entities, integrate access control checks at the service or domain layer, use security-focused code reviews to inspect multi-tenant logic, and apply automated checks for mass assignment and IDOR patterns in CI pipelines.

Find out what's already
exploitable in your codebase.