CVE-2026-105752
Last modified
CVE-2026-105752 is a low-severity vulnerability rated 3.1/10 on the CVSS scale. vLLM is an inference and serving engine for large language models. Prior to 0.30.0, Harmony tool continuations submitted through "POST /v1/responses" requests rebuild the next-turn engine input without preserving the cache_salt value, placing the continuation prefix in the global unsalted cache namespace even when the caller enabled salting. EPSS estimates a 0.21% chance of exploitation in the next 30 days.
Description
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, Harmony tool continuations submitted through "POST /v1/responses" requests rebuild the next-turn engine input without preserving the cache_salt value, placing the continuation prefix in the global unsalted cache namespace even when the caller enabled salting. On deployments with prefix caching enabled, which is the default, an authenticated tenant who can reconstruct a victim's low-entropy post-tool history can submit the same continuation and use the cached_tokens_per_turn count to determine whether the prefix was previously processed, defeating the intended tenant isolation of salted prefix caching. This issue is fixed in version 0.30.0.
Metrics
Weakness Enumeration
Affected Software
Source: CNA advisory (CVE.org). NVD analysis pending.
| Vendor | Product | Versions |
|---|---|---|
| vllm-project | vllm | < 0.30.0 |
References
Timeline
- Published
- Last Modified
- Status
- Awaiting Analysis
Frequently Asked Questions
What is CVE-2026-105752?
How severe is CVE-2026-105752?
How do I fix CVE-2026-105752?
How Strix Helps
- How Strix found a critical auth bypass in etcdStrix autonomously discovered a critical authentication bypass in etcd, later designated CVE-2026-33413.
- Autonomous PentestingAI agents that find and validate exploitable vulnerabilities like this one across your applications.
- PR ReviewsPentest every pull request so vulnerable code is caught before it ships to production.
- AI Penetration TestingHow AI-driven penetration testing continuously covers your attack surface.
Related CVEs from 2026
- CVE-2026-105747Docling simplifies document processing by parsing diverse fo…4.3
- CVE-2026-105748Docling simplifies document processing by parsing diverse fo…4.3
- CVE-2026-105749Docling simplifies document processing by parsing diverse fo…6.5
- CVE-2026-10575IBM MQ could allow an authenticated attacker to cause a deni…8.8
- CVE-2026-105750Docling simplifies document processing by parsing diverse fo…5.9
- CVE-2026-105751Docling simplifies document processing by parsing diverse fo…6.9
- CVE-2026-105753vLLM is an inference and serving engine for large language m…6.5
- CVE-2026-105754vLLM is an inference and serving engine for large language m…6.5
- CVE-2026-105755vLLM is an inference and serving engine for large language m…4.2
- CVE-2026-105756vLLM is an inference and serving engine for large language m…6.5
- CVE-2026-105757vLLM is an inference and serving engine for large language m…6.5
- CVE-2026-105758vLLM is an inference and serving engine for large language m…5.3
Are you affected by CVE-2026-105752?
Run a free Strix scan to check your systems for this vulnerability.
Scan your code nowSource: NVD / NIST
