CVE-2025-25183: vLLM prefix caching hash collisions can reuse cache from other prompts

Summary

CVE-2025-25183 affects vLLM, a high-throughput inference and serving engine for LLMs. Maliciously constructed statements can cause hash collisions in prefix caching, which relies on Python's built-in hash() function, so cached content generated from different prompts can be reused and interfere with later responses. Because hash(None) became a predictable constant in Python 3.12, an attacker who knows the prompts in use could populate the cache with a colliding prompt.

Mitigation

Fixed in version 0.7.2; all users are advised to upgrade. There are no known workarounds for this vulnerability.

Affected packages in the Exposure Registry

Matched by package name and ecosystem. Each entry shows whether the package delegates to a language model and how many tracked packages depend on it.

• vllmPyPILLM dependency since 2023-08-25 · 15 tracked dependents

Topics

• Inference infrastructure

Related items

• MediumCVE-2026-121936: easyVoice server-side request forgery through openaiBaseUrl in /api/v1/ttsSimilar attack · NVD/CVE Database
• MediumCVE-2026-118628: MLflow get-artifacts server-side request forgery information disclosureSimilar attack · NVD/CVE Database
• HighCVE-2026-118625: MLflow Tracking Server privilege escalation via get_endpointsSimilar attack · NVD/CVE Database
• InfoWhen trust becomes the payload: Inside the threats, etc quarterly reportSimilar attack · Elastic Security Labs
• MediumCritical Langflow flaws enable authentication bypass and RCESimilar attack · CSO Online