FastGPT’s code‑sandbox lacks stringent OS‑level isolation and relies solely on an application‑level 500ms polling interval to constrain memory usage. Attackers can exploit this architectural weakness by timing attacks that bypass the memory checks or by flooding the system with concurrent CPU‑intensive requests, thereby exhausting the JavaScript worker pool. The immediate consequence is a complete denial of service for legitimate users.

The vulnerability affects the labring FastGPT platform in all releases 4.14.13 and earlier. Users running these versions on any operating system where the sandbox is active are exposed to the issue.

With a CVSS score of 6.3 the risk is moderate. No EPSS data is available, and the vulnerability is not listed in CISA’s KEV catalog, indicating that it has not yet been widely exploited in the wild. However, the lack of any OS‑level resource limits means that an attacker can readily launch the described attacks without needing elevated credentials, rendering the exploit straightforward and scalable. The moderate severity combined with the potential for full service disruption underscores the need for timely mitigations.