/

CVE-2026-32995

Severity:

HIGH RISK

(7.5)

Broken Access Control / Authorization in Rocket.Chat DDP API

CVE-2026-32995

Awaiting Analysis

Vulnerability type

Category:

Broken Access Control / Authorization

CWE:

CWE-284

Impact:

Unauthorized Message Disclosure / Privacy Violation

Severity:

HIGH RISK

(7.5)

CVSS version:

3.0

Catch this CVE while code is being written

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

Summary

The Rocket.Chat DDP method autoTranslate.translateMessage incorrectly accepts a client-supplied IMessage object and forwards it to translateMessage() without checking the caller's identity or room membership.

As a result, any authenticated DDP user can retrieve the contents of arbitrary messages by ID from private channels, direct messages, and even E2EE rooms, bypassing normal access controls.

Affected context

This vulnerability occurs in Rocket.Chat deployments that expose the DDP API and use the autoTranslate.translateMessage method for translating message content.

The issue surfaces when the server trusts a client-supplied IMessage object and performs translation without confirming that:

  • The caller is a valid user for the current session, and
  • The user is a member of the room where the target message resides.

In such configurations, any logged-in DDP client can leverage this method to access messages from rooms they should not be able to read.

Exploit preconditions

  • Attack surface: Rocket.Chat DDP API, specifically the autoTranslate.translateMessage method.
  • Required access: Any authenticated Rocket.Chat user account with the ability to connect via DDP.
  • Trigger point: Issuing a DDP call to autoTranslate.translateMessage while supplying an IMessage object containing the ID of a target message.
  • Why exploitation is possible: The server method:
  • Does not verify Meteor.userId() appropriately for the operation context.
  • Does not confirm that the caller is a member of the room where the message is stored.
  • Accepts message identifiers and objects from the client as trustworthy, enabling direct access to message content based solely on ID.

Why this happens

The underlying flaw is an authorization bypass caused by missing or incomplete access control checks in a server-side DDP method.

The method autoTranslate.translateMessage accepts a client-supplied IMessage object and passes it directly to translateMessage() without:

  • Enforcing a permission check bound to the current session user, and
  • Verifying room membership or equivalent authorization for the referenced message.

This maps to CWE-284: Improper Access Control, where sensitive operations are performed without validating that the caller is entitled to perform them.

Conceptually, the bug arises from treating client-provided message identifiers as trusted handles to internal data, rather than requiring the server to resolve messages based on the authenticated user's visible context.

Potential impact

  • Unauthorized reading of private channel messages by any authenticated user.
  • Exposure of confidential content contained in direct messages between other users.
  • Disclosure of message content in E2EE rooms if translation occurs on the server-side representation of those messages.
  • Leakage of sensitive personal, operational, or security-related information shared in supposedly restricted rooms.
  • Potential regulatory or contractual non-compliance due to privacy and data protection violations.

General mitigation guidance

Effective mitigation focuses on enforcing strict, server-side authorization checks and minimizing trust in client-submitted identifiers:

  • Ensure every server-side method that accesses message content validates the authenticated user (e.g., via Meteor.userId() or equivalent session binding).
  • Perform room membership and permission checks server-side before reading or transforming any message content, regardless of client-provided parameters.
  • Avoid accepting complete message objects from the client; instead, accept minimal identifiers and have the server retrieve and validate resources based on the active user's allowed scope.
  • Apply a "default deny" policy for sensitive operations (such as message translation) so that lack of explicit authorization fails closed.
  • Reduce privileges of generic actions and separate translation capability from direct message read operations where possible.

These patterns help ensure that only users with legitimate access to a room can interact with its messages through auxiliary features like translation.

How teams detect issues like this

Organizations can approach detection by examining both code and runtime behavior:

  • Static analysis and code review:
  • Search for DDP or RPC methods that accept resource IDs or objects from the client and then call lower-level APIs without explicit authorization checks.
  • Flag sites where Meteor.userId() (or equivalent identity checks) is absent in methods that handle message, room, or user data.
  • Review helper functions like translateMessage() to ensure they are not exposed to untrusted callers without an authorization layer.
  • Policy and design checks:
  • Adopt guidelines that server methods must derive accessible resources from the authenticated user's context, not from client-supplied objects.
  • Use architectural reviews to identify "side-channel" features (translation, search, exporting, analytics) that may bypass primary access control logic.
  • Runtime monitoring:
  • Monitor for abnormal DDP calls that reference messages from many different rooms or high volumes of message IDs from a single user.
  • Employ anomaly detection on API usage patterns to catch mass access attempts to private messages.

Automated code scanning and CI-integrated review tools can help enforce these patterns across the codebase.

Sources

SourceLink
github.comgithub.com/RocketChat/Rocket.Chat/pull/40528
hackerone.comhackerone.com/reports/3734326

Frequently asked

questions

Yes. Any user who can authenticate to a vulnerable Rocket.Chat instance and access the DDP API can exploit this issue over the network, subject to the deployment's external exposure of DDP.

No elevated privileges are required. A standard authenticated account is sufficient, as the flawed method fails to verify room membership or enforce proper authorization.

The robust fix is to enforce server-side authorization checks whenever accessing message content: validate the authenticated user, confirm room membership or equivalent permissions for the referenced message, and avoid trusting client-supplied message objects. If the user does not have read access to the room, the method should deny the request.

Prevention hinges on a consistent access-control strategy: centralize authorization logic, require explicit checks in every server method that reads or manipulates messages, prohibit trusting client-side object data for sensitive resources, and include security-focused code review and automated checks that flag methods interacting with messages or rooms without clear permission validations.

Find out what's already
exploitable in your codebase.