CVE-2023-49471 is a high severity blind server-side request forgery, scored 8.8, in karlomikus Bar Assistant before 3.2.0. An authenticated user can make the server send requests. Upgrade to 3.2.0 or later and confirm the fix in the project's release notes.
What's affected
The record says Bar Assistant before version 3.2.0 does not validate a parameter before making a request through Image::make(), and that this could let authenticated remote attackers execute arbitrary code. The NVD vector is CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H.
The only reference is a researcher page. It describes the issue as a blind SSRF that logged in users reach through an image upload feature that takes a URL, and it lists the affected product as before 3.2.0. It does not state a fixed version explicitly, which is why the project's release notes should confirm that 3.2.0 contains the fix. Note that the record's code execution claim is stronger than the SSRF description on the page. I only know that the record says so.
How to check if you're exposed
Find the running version in the application's about or settings page, or from the image tag you deploy.
If you run it in Docker, list the image tags in use.
docker ps --format '{{.Image}}'
Any tag below 3.2.0 is in the affected range. Then check the project's release notes for 3.2.0 to confirm the fix.
What to do now
- Find your Bar Assistant version and compare it with 3.2.0.
- Upgrade to 3.2.0 or later and read the release notes to confirm the fix.
- Review which accounts can upload images, and remove accounts you do not recognize, since the flaw needs a login.
- Limit outbound network access from the host so it cannot reach internal services or cloud metadata addresses.
- Check access logs for unusual image URLs submitted by users.
Why it matters
The weakness is server-side request forgery, CWE-918. The server fetches a URL that a user supplied, so the user decides where the server connects, including addresses an outsider cannot reach directly.
Blind means the response is not shown to the attacker, but internal services can still be probed or poked. That is why egress filtering is a good second layer for any application that fetches user supplied URLs. KEV status is false for this CVE.
Try it yourself
Click through it at your own pace, or open it full screen.
Quick check
What does blind SSRF mean?
The server makes the request, but the caller does not see the response. The attacker can still cause actions or learn from side effects such as timing.
Sources
Researcher page referenced by NVD
Last reviewed: 19 October 2026.
