CVE-2025-48944: vLLM Tool Schema Allows DoS via Malformed Fields

# CVE-2025-48944: vLLM Tool Schema Allows DoS via Malformed Fields

https://gcdradar.vercel.app/incidents/cve-2025-48944-vllm-tool-schema-allows-dos-via-malformed-fie · AIR-2025-0286 · closed 2025-05-28

- Lane: Patch now
- Severity: 2 (Disclosed vulnerability)
- Fix status: Fix status unknown
- First reported: 2025-05-28
- AI role: Attack on AI
- Products: vLLM
- CVE ids: CVE-2025-48944
- CVSS: 6.5

## Summary

The vLLM backend endpoint lacks input validation for pattern and type fields during tool invocations. Malformed requests can crash inference workers, causing service downtime until manual restarts occur.

## What to do

No action recorded; read the sources below and the vendor's advisory.

## Latest updates

- 2025-05-30: CVE-2025-48944: vLLM is an inference and serving engine for large language models (LLMs). (https://nvd.nist.gov/vuln/detail/CVE-2025-48944)
- 2025-05-28: CVE-2025-48944: vLLM Tool Schema allows DoS via Malformed pattern and type Fields (https://github.com/advisories/GHSA-vrq3-r879-7m65)

## Sources

- 2025-05-28 GitHub Advisory Database: CVE-2025-48944: vLLM Tool Schema allows DoS via Malformed pattern and type Fields https://github.com/advisories/GHSA-vrq3-r879-7m65
- 2025-05-30 NVD: CVE-2025-48944: vLLM is an inference and serving engine for large language models (LLMs). https://nvd.nist.gov/vuln/detail/CVE-2025-48944

Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.