Is CVE-2026-85704 exploited in the wild?

No. Junglewise Threat Intelligence has no report of CVE-2026-85704 being exploited in the wild.

How severe is CVE-2026-85704?

Low severity, with a CVSS score of 3.7, published 4 September 2026.

Which products and versions are affected?

Ramon-Victor Freegpt-Webui up to 098db3dfeb41555c2ca9269df0f13e10ec1c35dc (rolling release).

Is there a fix for CVE-2026-85704?

No fixed version is recorded yet. Check the vendor's advisory in the references below. If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.

How is CVE-2026-85704 attacked?

It is attacked over the network.

Am I affected by CVE-2026-85704?

Check whether you run Ramon-Victor Freegpt-Webui at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

CVE-2026-85704: ramon-victor freegpt-webui race condition in jailbreak mode

CVE-2026-85704 · Severity: low · CVSS 3.7 · Published 2026-09-04

freegpt-webui is a web application that provides access to GPT models. A race condition in the jailbreak mode feature allows an attacker to send concurrent requests that cause unbounded prompt accumulation, eventually exceeding token limits and causing a denial of service affecting all users. The vulnerability requires high attack complexity and is in an archived, unsupported product.

Technologies: Ramon-Victor Freegpt-Webui. Vendors: Ramon-Victor.

Quick answers

Is CVE-2026-85704 exploited in the wild?: No. Junglewise Threat Intelligence has no report of CVE-2026-85704 being exploited in the wild.
How severe is CVE-2026-85704?: Low severity, with a CVSS score of 3.7, published 4 September 2026.
Which products and versions are affected?: Ramon-Victor Freegpt-Webui up to 098db3dfeb41555c2ca9269df0f13e10ec1c35dc (rolling release).
Is there a fix for CVE-2026-85704?: No fixed version is recorded yet. Check the vendor's advisory in the references below. If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.
How is CVE-2026-85704 attacked?: It is attacked over the network.
Am I affected by CVE-2026-85704?: Check whether you run Ramon-Victor Freegpt-Webui at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

Technical details

The vulnerability is a CWE-362 race condition (TOCTOU) in the getJailbreak function of server/config.py. The function performs a non-atomic read-modify-write operation (+=) on a global, module-level dictionary (special_instructions) without synchronization. An attacker can send multiple concurrent POST requests with the jailbreak parameter to trigger interleaved reads and writes, causing the prompt instruction content to accumulate on each request. After sufficient iterations, the cumulative instruction text exceeds the upstream model's token limit, resulting in persistent denial of service for all users attempting to use the jailbreak feature. The attack is network-accessible but requires high complexity to exploit reliably.

Affected products

ramon-victor freegpt-webui up to 098db3dfeb41555c2ca9269df0f13e10ec1c35dc (rolling release)

Timeline

2026-07-18: disclosed: Security advisory published on GitHub gist
2026-09-04: advisory: CVE-2026-85704 published on NVD

References

This page as Markdown. Free to reuse under CC BY 4.0 with a link back.