low
2.6
CWE
354
EPSS
0.045%
Advisory Published
CVE Published
Updated

CVE-2025-25183: vLLM using built-in hash() from Python 3.12 leads to predictable hash collisions in vLLM prefix cache

First published: Thu Feb 06 2025(Updated: )

### Summary Maliciously constructed prompts can lead to hash collisions, resulting in prefix cache reuse, which can interfere with subsequent responses and cause unintended behavior. ### Details vLLM's 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. ### Impact 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. ### Solution We address this problem by initializing hashes in vllm with a value that is no longer constant and predictable. It will be different each time vllm runs. This restores behavior we got in Python versions prior to 3.12. Using a hashing algorithm that is less prone to collision (like sha256, for example) would be the best way to avoid the possibility of a collision. However, it would have an impact to both performance and memory footprint. Hash collisions may still occur, though they are no longer straight forward to predict. To give an idea of the likelihood of a collision, for randomly generated hash values (assuming the hash generation built into Python is uniformly distributed), with a cache capacity of 50,000 messages and an average prompt length of 300, a collision will occur on average once every 1 trillion requests. ### References * https://github.com/vllm-project/vllm/pull/12621 * https://github.com/python/cpython/commit/432117cd1f59c76d97da2eaff55a7d758301dbc7 * https://github.com/python/cpython/pull/99541

Credit: security-advisories@github.com

Affected SoftwareAffected VersionHow to fix
pip/vllm<0.7.2
0.7.2
vLLM<0.7.2

Never miss a vulnerability like this again

Sign up to SecAlerts for real-time vulnerability data matched to your software, aggregated from hundreds of sources.

Read MoreStart Free Trial

Reference Links

Frequently Asked Questions

  • What is the severity of CVE-2025-25183?

    CVE-2025-25183 is considered a moderate severity vulnerability due to its potential impact on the functionality of affected systems.

  • How do I fix CVE-2025-25183?

    To fix CVE-2025-25183, upgrade to vLLM version 0.7.3 or later.

  • What type of attacks can be executed through CVE-2025-25183?

    CVE-2025-25183 allows for hash collision attacks that cause cache reuse, leading to interference in subsequent responses.

  • Which versions of vLLM are affected by CVE-2025-25183?

    CVE-2025-25183 affects all versions of vLLM up to and including version 0.7.2.

  • Is CVE-2025-25183 related to any specific components of vLLM?

    CVE-2025-25183 specifically impacts the prefix caching mechanism within the vLLM inference engine.

Company

Resources

Contact

SecAlerts Pty Ltd.
132 Wickham Terrace
Fortitude Valley,
QLD 4006, Australia
info@secalerts.co
By using SecAlerts services, you agree to our services end-user license agreement. This website is safeguarded by reCAPTCHA and governed by the Google Privacy Policy and Terms of Service. All names, logos, and brands of products are owned by their respective owners, and any usage of these names, logos, and brands for identification purposes only does not imply endorsement. If you possess any content that requires removal, please get in touch with us.
© 2025 SecAlerts Pty Ltd.
ABN: 70 645 966 203, ACN: 645 966 203