CVE Watchtower

CVE-2026-104885io.openremote:openremote-manager Missing Authorization

NVD

Vulnerability Summary

### Summary

Cross-realm MQTT session disconnect without authorization

## Vulnerability

The `disconnectUserSession` endpoint in `UserResourceImpl.java` (line 443) accepts a `realm` path parameter but ignores it entirely, performing no authorization or realm-boundary check before force-disconnecting an MQTT session. Any authenticated user from any realm can disconnect MQTT broker sessions belonging to users in other realms by supplying the target session ID.

File: `manager/src/main/java/org/openremote/manager/security/UserResourceImpl.java`, lines 443-447

```java
@Override
public void disconnectUserSession(RequestParams requestParams, String realm, String sessionID) {
if (!mqttBrokerService.disconnectSession(sessionID)) {
throw new NotFoundException("User session not found");
}
}
```

The REST interface at `model/src/main/java/org/openremote/model/security/UserResource.java` (line 201-204) maps this to `GET {realm}/disconnect/{sessionID}` with no `@RolesAllowed` annotation, so any authenticated user (not just admins) can call it.

## Root Cause

CVE-2026-54641 (commit de89b8d) fixed cross-realm information disclosure in three read methods of the same file (`get()`, `getUserClientRoles()`, `getUserRealmRoles()`) and also added a realm check to `getClientRoles()`. The fix added `throwIfCannotReadUserInRealm()` and `throwIfCannotAdminRealm()` guards. However, `disconnectUserSession` was not patched and remains unprotected. Every other mutating method in this class has either a `@RolesAllowed` annotation on the interface or an explicit realm check in the implementation. This one has neither.

## Proof of Concept

Given two realms, "master" and "building":

1. User A authenticates in the "building" realm (any role, no admin needed).
2. User B has an active MQTT session in the "master" realm with session ID `S`.
3. User A sends: `GET /api/master/user/master/disconnect/S` with their building-realm bearer token.
4. The handler ignores the realm path segments, looks up session `S` in the global `connectionIDConnectionMap`, and calls `doForceDisconnect()`. User B's MQTT connection is terminated.

Session IDs use the Netty `ChannelId.asLongText()` format (deterministic, partially sequential). While not trivially brute-forceable, IDs can leak through monitoring, logs, or correlated API calls. Within the same realm, any user (not just admins) can also disconnect sessions they should not have access to.

## Impact

In a multi-tenant IoT deployment, an attacker authenticated in any realm can disrupt MQTT device connectivity in other realms. Since OpenRemote is used for building management, energy monitoring, and industrial IoT, forcibly disconnecting MQTT sessions can:

- Knock IoT sensors and actuators offline across tenant boundaries
- Disrupt safety-critical alarm monitoring (compounding the alarm IDOR fixed in CVE-2026-57168)
- Cause denial of service against other tenants' device fleets

The method also uses HTTP GET for a destructive operation, making it susceptible to CSRF via image tags or link prefetch.

## Suggested Fix

Add a realm administration check consistent with the pattern used elsewhere in this file:

```java
@Override
public void disconnectUserSession(RequestParams requestParams, String realm, String sessionID) {
throwIfCannotAdminRealm(realm);

RemotingConnection connection = mqttBrokerService.getConnectionByID(sessionID);
if (connection == null) {
throw new NotFoundException("User session not found");
}

// Verify the session belongs to a user in the requested realm
// before disconnecting

if (!mqttBrokerService.disconnectSession(sessionID)) {
throw new NotFoundException("User session not found");
}
}
```

At minimum, add `throwIfCannotAdminRealm(realm)` as the first line, and consider adding a `@RolesAllowed(Constants.WRITE_ADMIN_ROLE)` annotation to the interface method. The HTTP method should also be changed from GET to DELETE.

### AI tooling

AI assistance was used for the code audit and for drafting this report. The finding against the project's source at the location cited above was manually verified before reporting it, and the severity and impact assessment are the reporter's own.
Severity Level
MEDIUM(6.3)
Published Date
Nov 19, 2026
Last Modified
Nov 20, 2026
Exploitation Status
Public exploit available
CVE Record Status
Published
EPSS Score (30-Day)
0.47%Probability
Root Weakness (CWE)
The software does not perform an authorization check when an actor attempts to access a resource or perform an action.
CVSS v4.0 Base Metrics — Score 6.3
CVSS:4.0/AV:N/AC:H/AT:N/PR:N/UI:N/VC:N/VI:N/VA:L/SC:N/SI:N/SA:N
Attack VectorNetwork
Attack ComplexityHigh
Attack RequirementsNone
Privileges RequiredNone
User InteractionNone
Confidentiality (Vulnerable System)None
Integrity (Vulnerable System)None
Availability (Vulnerable System)Low
Confidentiality (Subsequent System)None
Integrity (Subsequent System)None
Availability (Subsequent System)None

Affected & Patched Versions

Affected Versions
  • io.openremote:openremote-manager < 1.31.0
Patched Versions
No patched version published yet
Until a fix ships, reduce exposure:
  • Restrict access to the affected service or interface to trusted networks.
  • Disable the affected feature if you do not need it.
  • Watch the vendor advisory for a firmware or software update.

What to do

Prioritize this fix. Exploit code is public, so attacks can follow quickly.

  1. Check whether you run io.openremote:openremote-manager and which version.
  2. No patched version is listed yet. Limit access to the affected service to trusted networks, and disable the feature if you can.
  3. Subscribe to the vendor advisory so you learn when a fix ships.
  4. Review logs and exposed services for signs of earlier compromise.

Timeline

  • Nov 19, 2026Published
  • Nov 20, 2026Record last modified

Score, severity and exploitation changes are tracked from the day Watchtower started logging them.

Get notified about new io.openremote:openremote-manager CVEsYou are reading CVE-2026-104885. Get an email when KEV listings, exploits or new CVEs match your stack.
Set up CVE alerts

Data last checked: Nov 20, 2026 17:00 UTC