CVE-2026-32128 FastGPT Python Sandbox Bypass of File-Write Restriction

What is CVE-2026-32128?

CVE-2026-32128 is a medium-severity vulnerability in labring FastGPT, published on 11 Mar 2026. Any signed-in user can exploit it over the network, with no action from the victim.

Official description

FastGPT is an AI Agent building platform. In 4.14.7 and earlier, FastGPT's Python Sandbox (fastgpt-sandbox) includes guardrails intended to prevent file writes (static detection + seccomp). These guardrails are bypassable by remapping stdout (fd 1) to an arbitrary writable file descriptor using fcntl. After remapping, writing via sys.stdout.write() still satisfies the seccomp rule write(fd==1), enabling arbitrary file creation/overwrite inside the sandbox container despite the intended no file writes restriction.

Weakness: CWE-184

Which versions are affected by CVE-2026-32128?

It affects 1 product.

Products include Fastgpt Fastgpt.

How do I fix CVE-2026-32128?

No fix has been published for CVE-2026-32128 yet. Limit exposure of the affected software until one is available.

Is CVE-2026-32128 being exploited?

It is not confirmed as exploited in the wild, but public exploit code for CVE-2026-32128 exists.

Public exploit code: 1 proofs of concept on GitHub.

Other vulnerabilities in labring FastGPT

• CVE-2026-34162 FastGPT: Unauthenticated SSRF via httpTools Endpoint Leads to Internal API Key Theft Critical 10.0
• CVE-2026-40351 FastGPT: NoSQL Injection in loginByPassword leads to Authentication Bypass Critical 9.8
• CVE-2026-33075 FastGPT has Arbitrary Code Execution in GitHub Actions via pull_request_target in fastgpt-preview-image.yml Critical 9.4
• CVE-2026-40352 FastGPT: NoSQL Injection in updatePasswordByOld Leads to Account Takeover High 8.8
• CVE-2026-34163 Server-Side Request Forgery via MCP Tools Endpoint in FastGPT High 7.7

Sources

• CVE record
• NVD
• OSV
• GitHub advisory
• How this page is made

Also known as GHSA-6hw6-mxrm-v6wj.