Cache poisoning via incomplete multimodal hashing
Published May 29, 2025 · Updated May 29, 2025
Hash collision in vLLM 0.7.0 through 0.8.5.post1 allows remote authenticated users to trigger incorrect multimodal cache hits. MultiModalHasher.serialize_item hashes only raw image pixels and flattened tensor or array bytes, omitting dimensions, mode, dtype, and shape that distinguish inputs. Low-privileged network access and a deliberately colliding image or video input are required; a cache hit can return an abnormal response or disclose cached data.
Summary
What happened
Hash collision in vLLM 0.7.0 through 0.8.5.post1 allows remote authenticated users to trigger incorrect multimodal cache hits. MultiModalHasher.serialize_item hashes only raw image pixels and flattened tensor or array bytes, omitting dimensions, mode, dtype, and shape that distinguish inputs. Low-privileged network access and a deliberately colliding image or video input are required; a cache hit can return an abnormal response or disclose cached data.
The record
- CVE
- CVE-2025-46722
- Published
- May 29, 2025
- Updated
- May 29, 2025
- Vendor
- vLLM Project
- Product
- vLLM
- Classifications
- CWE-1023, CWE-1288, T1190
- Attack vector
- network
- Privileges
- authenticated
Timeline
How it unfolded
- May 29, 2025CVE publishedPublication date reported by the CVE source.
- May 29, 2025Record updatedLatest update available in the CVE record.
Exploitability
Present is not the same as exploitable
Compare your product and version with the public record. A matching version still requires validation against your environment.
Is a vulnerable build present?
Compare these published version ranges with your installed build and any vendor patches.
- Affected versionversion=>= 0.7.0, < 0.9.0
What conditions does exploitation require?
What is affected?
Published CVSS scores
CVSS describes severity. EPSS estimates exploitation probability.
Attacks
What attackers are doing with it
Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.
Weakness, pattern, technique
Public exploit references
No public exploit references are available in this record.
Labels summarize the accepted research assessment. They do not indicate a test against your environment.
Technologies
Your stack
See the directory against your own environment.
Your stack
Check the software in your environment
Book a demo to see how Hinoki identifies affected software and validates exploitability in your environment.
Book a demo