FastGPT versions from 4.14.10 to before 4.14.13 contain a missing authentication vulnerability (CWE-306) in the agent-sandbox component. The startup script entrypoint.sh configures code-server with the --auth none flag and binds it to 0.0.0.0:8080, exposing the service without authentication to any network user. This misconfiguration allows unauthenticated remote code execution (RCE) on the sandbox environment. The vulnerability has a CVSS 3.1 score of 9.8, indicating critical severity. The vendor patched this issue in version 4.14.13.

An attacker with network access to port 8080 on affected FastGPT instances can bypass authentication entirely and execute arbitrary code remotely within the sandbox environment. This leads to full compromise of the sandbox, including confidentiality, integrity, and availability impacts. The vulnerability is critical due to the ease of exploitation (no privileges or user interaction required) and the high impact on the system.

Upgrade FastGPT to version 4.14.13 or later, where this vulnerability has been patched by removing the insecure --auth none configuration and restricting network exposure. Until the upgrade is applied, restrict network access to port 8080 to trusted users only to reduce exposure. Patch status is confirmed by the vendor's versioning information.