/

CVE-2026-41190

Severity:

HIGH RISK

(7.1)

Missing Authorization in FreeScout Draft Saving Endpoint

CVE-2026-41190

Deferred

Vulnerability type

Category:

Missing Authorization

CWE:

CWE-863

Impact:

Unauthorized Draft Creation and Conversation Manipulation

Severity:

HIGH RISK

(7.1)

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

FreeScout versions prior to 1.8.215 contain a missing authorization check in the save_draft AJAX endpoint when APP_SHOW_ONLY_ASSIGNED_CONVERSATIONS is enabled.

An authenticated user who is neither the assignee nor creator of a conversation can directly POST to this endpoint and create drafts in conversations that are hidden from them in the UI, potentially manipulating or polluting ticket content.

Affected context

This vulnerability affects FreeScout deployments that:

  • Enable the APP_SHOW_ONLY_ASSIGNED_CONVERSATIONS configuration, and
  • Rely on role- and assignment-based restrictions to limit who can interact with particular conversations.

In this context, the main conversation view correctly hides conversations from non-assigned users, but the save_draft AJAX path does not consistently enforce the same assignment and ownership checks.

Exploit preconditions

  • Attack surface: Application HTTP interface, specifically the AJAX save_draft endpoint.
  • Required access: Authenticated FreeScout account with access to the help desk interface, but without assignment or ownership of the target conversation.
  • Trigger point: Directly sending a crafted POST request to the save_draft endpoint referencing a conversation that is not visible in the UI.
  • Why exploitation is possible: The endpoint does not reliably verify that the caller is the assignee or creator of the conversation before accepting and storing a draft, leading to a broken authorization decision (CWE-863).

Why this happens

The core issue is an inconsistent authorization model between the conversation viewing logic and the save_draft AJAX handler.

While the UI-level conversation view checks whether a user is an assignee or creator before displaying a conversation, the draft-saving path assumes that any authenticated user who can reach the endpoint is authorized to act on the referenced conversation ID.

This pattern maps to CWE-863: Incorrect Authorization, where sensitive operations are exposed without enforcing the same access control rules that protect related resources.

Potential impact

  • Unauthorized users may create or update drafts in conversations they cannot normally see.
  • Conversation history could be polluted with misleading or unwanted draft content.
  • Internal workflows may be disrupted if agents rely on drafts as part of their handling process.
  • Information about the existence or identifiers of hidden conversations may be inferred from endpoint behavior and responses.

General mitigation guidance

Mitigation generally focuses on aligning all backend endpoints with a centralized authorization model:

  • Apply consistent authorization checks for every action that references a conversation, ensuring the caller is permitted to access and modify that resource.
  • Route all conversation-related operations (viewing, drafting, replying, tagging) through common authorization helpers that implement assignment and ownership rules.
  • Design APIs so that they derive the target resource scope from the authenticated user's context wherever possible, rather than trusting arbitrary IDs sent by clients.
  • Use defense-in-depth measures by limiting what low-privilege accounts can do, even when they can reach internal AJAX endpoints.

How teams detect issues like this

Organizations can identify similar authorization gaps by:

  • Reviewing controller and AJAX endpoint code for consistent use of authorization checks on conversation IDs and related objects.
  • Employing static analysis or custom linters to flag routes that perform state-changing operations without explicit permission checks.
  • Adding integration tests that simulate users with different roles and assignments attempting to perform cross-conversation actions (e.g., saving drafts on unassigned tickets).
  • Monitoring logs for unusual patterns, such as draft creation events referencing conversations not normally visible to the acting user.

Sources

SourceLink
github.comgithub.com/freescout-help-desk/freescout/commit/414878eb79be7cb01a3ae124df6efcd23729275f
github.comgithub.com/freescout-help-desk/freescout/releases/tag/1.8.215
github.comgithub.com/freescout-help-desk/freescout/security/advisories/GHSA-vj2p-2789-3747

Frequently asked

questions

It is exploitable over the network but requires a valid FreeScout account; anonymous users cannot typically exploit it.

An attacker needs an authenticated user account on the affected FreeScout instance, but does not need to be assigned to or own the target conversation.

The most robust fix is to enforce the same authorization checks on the save_draft endpoint as on the conversation view, using a central authorization layer that validates assignment, ownership, and role before permitting any state changes.

Implement a unified access control API for all conversation-related operations, require explicit permission checks on every state-changing route, and back this with automated tests and code review practices that treat any endpoint accepting resource IDs from clients as high-risk.

Find out what's already
exploitable in your codebase.