CVE-2025-25183

Description

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Maliciously constructed statements can lead to hash collisions, resulting in cache reuse, which can interfere with subsequent responses and cause unintended behavior. Prefix caching makes use of Python's built-in hash() function. As of Python 3.12, the behavior of hash(None) has changed to be a predictable constant value. This makes it more feasible that someone could try exploit hash collisions. The impact of a collision would be using cache that was generated using different content. Given knowledge of prompts in use and predictable hashing behavior, someone could intentionally populate the cache using a prompt known to collide with another prompt in use. This issue has been addressed in version 0.7.2 and all users are advised to upgrade. There are no known workarounds for this vulnerability.

CVSS v3.1 Score

2.6

LOW

CVSS:3.1/AV:N/AC:H/PR:L/UI:R/S:U/C:N/I:L/A:N

Weakness Type (CWE)

CWE-354 CWE-354

Affected Products

Vendor	Product	Versions
vllm	vllm	< 0.7.2

References

Advisories & Patches

https://github.com/vllm-project/vllm/security/advisories/GHSA-rm76-4mrf-v9r8

Other References

https://github.com/python/cpython/commit/432117cd1f59c76d97da2eaff55a7d758301dbc7 https://github.com/vllm-project/vllm/pull/12621

Frequently Asked Questions

What is CVE-2025-25183? +

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Maliciously constructed statements can lead to hash collisions, resulting in cache reuse, which can interfere with subsequent responses and cause unintended behavior. Prefix caching makes use of Python's built-in hash() function. As of Python 3.12, the behavior of hash(None) has changed to be a predictable constant value. This makes it more feasible that someone could try exploit hash collisions. The impact of a collision would be using cache that was generated using different content. Given knowledge of prompts in use and predictable hashing behavior, someone could intentionally populate the cache using a prompt known to collide with another prompt in use. This issue has been addressed in version 0.7.2 and all users are advised to upgrade. There are no known workarounds for this vulnerability. It has a CVSS v3.1 base score of 2.6 (LOW).

How severe is CVE-2025-25183? +

CVE-2025-25183 has a CVSS v3.1 score of 2.6 out of 10, rated LOW. This is a low-severity vulnerability with limited impact.

What products are affected by CVE-2025-25183? +

CVE-2025-25183 affects products from vllm, specifically: vllm. Check the affected products table above for specific version ranges.

How do I check if I'm vulnerable to CVE-2025-25183? +

You can use Secably's free Website Scanner to check your website for known vulnerabilities. For infrastructure scanning, use the Port Scanner to identify exposed services that may be affected. Check the vendor advisories linked above for specific patch and version information.

Related Vulnerabilities

CVE-2025-11694

A security issue exists within 1769 CompactLogix controllers due to the missing validation of sequence numbers and source IP addresses …

CVE-2024-7402

Netskope has identified a potential gap in its agent (Netskope Client) in which a malicious insider can potentially tamper the …

CVE-2024-48930

secp256k1-node is a Node.js binding for an Optimized C library for EC operations on curve secp256k1. In `elliptic`-based version, `loadUncompressedPublicKey` …

CVE-2025-4616

An insufficient validation of an untrusted input vulnerability in Palo Alto Networks Prisma® Browser allows a locally authenticated non-admin user …

CVE-2024-25678 9.8

In LiteSpeed QUIC (LSQUIC) Library before 4.0.4, DCID validation is mishandled.

CVE-2025-11543 9.8

Improper Validation of Integrity Check Value vulnerability in Sharp Display Solutions projectors allows a attacker may create and run unauthorized …