Vulnerability record · CVE-2025-32375 · published 9 April 2025
CVE-2025-32375: BentoML runner server insecure deserialization enables unauthenticated RCE
Bentoml · Bentoml
BentoML before 1.4.8 deserializes untrusted data in its runner server. By sending a crafted POST request with specific headers and parameters, an attacker can execute arbitrary code on the server. The flaw is fixed in 1.4.8.
Description
BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.8, there was an insecure deserialization in BentoML's runner server. By setting specific headers and parameters in the POST request, it is possible to execute any unauthorized arbitrary code on the server, which will grant the attackers to have the initial access and information disclosure on the server. This vulnerability is fixed in 1.4.8.
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Automated analysis
critical priorityCVSS 9.8 with no authentication or interaction required, and public exploit code plus high EPSS make this an urgent remote code execution risk.
What it is
BentoML before 1.4.8 deserializes untrusted data in its runner server. By sending a crafted POST request with specific headers and parameters, an attacker can execute arbitrary code on the server. The flaw is fixed in 1.4.8.
Impact
An unauthenticated attacker gains remote code execution on the BentoML runner server, leading to initial access and information disclosure.
Attack surface
Reachable over the network via a POST request to the runner server; no authentication or user interaction is required per the CVSS vector (AV:N/AC:L/PR:N/UI:N).
Exploitation
Not listed in CISA KEV, but EPSS is 0.52415 (98.9th percentile) and the vendor advisory is tagged Exploit, indicating public exploit code exists.
What to do
- Upgrade BentoML to 1.4.8 or later immediately.
- Restrict network access to BentoML runner servers to trusted clients only.
- Place runner servers behind authentication and a reverse proxy where feasible.
- Monitor for and block POST requests containing unexpected serialized payloads or headers.
Detection
- Inspect BentoML runner server logs for anomalous POST requests with unusual headers or parameters.
- Monitor for unexpected child processes or outbound connections spawned by the BentoML process.
- Use network detection to flag serialized payload patterns in HTTP request bodies to runner endpoints.
This assessment is produced automatically and is not human-reviewed. Verify against the vendor advisory before acting on it.
Affected products
1 vulnerable configurations from NVD's CPE data, grouped by vendor and product.
References
| Link | Tags |
|---|---|
| https://github.com/bentoml/BentoML/security/advisories/GHSA-7v4r-c989-xh26 | ExploitVendor Advisory |
| https://github.com/bentoml/BentoML/security/advisories/GHSA-7v4r-c989-xh26 | ExploitVendor Advisory |
Track CVE-2025-32375 inside VULONE
Watch it alongside the ransomware crews, C2 infrastructure and forum chatter that reference it, query it through the API and pull it into your SIEM over TAXII.
Related vulnerabilities
Same products first, then exploited flaws of the same weakness class.
Source: NIST National Vulnerability Database (record CVE-2025-32375), CISA KEV, FIRST EPSS (scores of 2026-09-27). This page is refreshed as NVD updates the record.