CVE-2025-71379

Description

vLLM versions >= 0.6.3 and < 0.9.0 contain multiple regular expression denial of service (ReDoS) vulnerabilities. Several regex patterns — in vllm/lora/utils.py, the phi4mini tool parser, and the OpenAI-compatible serving chat endpoint — are susceptible to catastrophic backtracking. An attacker submitting crafted input with nested or repeated structures can trigger severe CPU consumption and performance degradation, resulting in denial of service.

Weakness classification

• disclosure@vulncheck.com: CWE-1333

Primary references

• disclosure@vulncheck.com — Exploit, Vendor Advisory
• disclosure@vulncheck.com — Third Party Advisory