CVE-2025-71379Medium· 4.3▾ SunlitvLLM 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 —…
▾ Sunlit zone — Low / medium · no exploitation signal
impact 23.7 · likelihood 0.1 · exploitation 0
Need a working PoC? Pro members can cast a request and our team develops one — it lands right here.
Exploit-prediction probability, daily snapshots since Aug 4.
Disclosure to exploitation, from the record and what we observed since indexing it.
Disclosed via NVD
0.3%
0.3% → 0.5%
Last analysed / modified upstream
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.
vllm >= 0.6.3, < 0.9.0Upgrade past the affected range:
vllm 0.9.0Affected packages:
vllm >= 0.6.3, < 0.9.0Patched in:
vllm 0.9.0Connected by shared product, vendor, weakness, or advisory.
CVE-2025-48887Medium· 6.5vLLM has a Regular Expression Denial of Service (ReDoS, Exponential Complexity) Vulnerability in `pythonic_tool_parser.py`
GHSA-vfm7-4h43-gp6mMedium· 4.3Duplicate Advisory: vLLM Vulnerable to Regular Expression Denial of Service
CVE-2026-73556Medium· 5.3vLLM is an inference and serving engine for large language models
CVE-2026-115911Medium· 6.5vLLM: DoS via unbounded chat_template Jinja2 rendering exhausts the renderer thread pool
CVE-2026-115912High· 7.5vLLM: Public multimodal SHM handle forgery reaches worker-side `pickle.loads()` with insecure serialization disabled
CVE-2026-115913Medium· 6.5vLLM: Phi-4 multimodal audio encoder materializes and retains full attention matrices: a single max-length audio OOM-crashes the vLLM engine