Description

A security flaw has been discovered in ramon-victor freegpt-webui up to 098db3dfeb41555c2ca9269df0f13e10ec1c35dc. This issue affects the function getJailbreak of the file server/config.py of the component Jailbreak Mode. The manipulation results in race condition. It is possible to launch the attack remotely. The attack requires a high level of complexity. The exploitability is assessed as difficult. The exploit has been released to the public and may be used for attacks. This product takes the approach of rolling releases to provide continious delivery. Therefore, version details for affected and updated releases are not available. This vulnerability only affects products that are no longer supported by the maintainer.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "3.7",
  "pubDate": "2026-09-04T21:17:26.863Z",
  "pubdate": "2026-09-04T21:17:26.863Z",
  "executiveSummary": "A critical race condition vulnerability has been identified in the 'getJailbreak' function within the server/config.py file of the ramon-victor freegpt-webui component.\nThis vulnerability allows remote attackers to exploit synchronization flaws during the configuration retrieval process.\nThe flaw affects all versions of the product up to commit 098db3dfeb41555c2ca9269df0f13e10ec1c35dc.\nThe risk is exacerbated by the availability of public exploits, increasing the likelihood of exploitation by threat actors.\nThe vulnerability is characterized by a high level of complexity and difficulty in exploitability; however, it remains a significant security risk for unsupported instances.\nImpacts include potential unauthorized modification or retrieval of sensitive jailbreak configurations, which may allow for the bypass of security controls intended to limit LLM output behavior.\nAs the software utilizes a rolling release model and the maintainer no longer supports these affected versions, manual remediation or migration is required for continued secure operation.",
  "technicalDetails": "The root cause of the vulnerability resides in the getJailbreak function within server/config.py, which fails to implement proper thread-safe synchronization or atomicity when handling jailbreak configuration states.\nThe race condition occurs when concurrent read and write operations access the internal state of the configuration object during the execution of getJailbreak. Because the system lacks adequate locking mechanisms (such as mutexes or read-write locks), an attacker can trigger a race between the loading of the configuration and the application of security constraints.\nThe attack flow requires the attacker to send a high volume of precisely timed, concurrent requests to the affected endpoint. By flooding the server with requests, the attacker aims to create a condition where the thread executing getJailbreak accesses an inconsistent or partially updated memory state. If the timing is successful, the application may inadvertently serve or apply a 'jailbroken' state before security filters or overrides are finalized.\nExploitation is inherently complex as it depends on the specific scheduling patterns of the underlying Python runtime and the server's load. However, because the exploit is publicly available, the barrier to entry for potential attackers is reduced. The exploitation does not strictly require high-level privileges, as the endpoint is typically exposed to facilitate user interaction with the jailbreak functionality.\nThe vulnerable component, Jailbreak Mode, is designed to sanitize or gate output; by manipulating the execution flow, the attacker effectively disables these safety mechanisms. Post-exploitation, the attacker can leverage the compromised configuration to induce the LLM to ignore safety guardrails, leading to the generation of prohibited or harmful content.\nThis vulnerability is particularly persistent in legacy deployments where the software has not been updated or where the maintainer has ceased development, leaving the system in a perpetually vulnerable state with no path for official security updates."
}