CVE-2026-111392

SGLang through 0.5.21 trusts caller-supplied mmhashes values as the multimodal prefix-cache key, allowing remote attackers to poison cached KV entries across requests. Attackers can send a different image under a known or predicted hash via /generate to disclose another request's image content or...

EUVD-2026-103024

SGLang through 0.5.21 trusts caller-supplied mmhashes values as the multimodal prefix-cache key, allowing remote attackers to poison cached KV entries across requests. Attackers can send a different image under a known or predicted hash via /generate to disclose another request's image content or...

PT-2026-116001

SGLang through 0.5.21 trusts caller-supplied mm hashes values as the multimodal prefix-cache key, allowing remote attackers to poison cached KV entries across requests. Attackers can send a different image under a known or predicted hash via /generate to disclose another request's image content o...

PYSEC-2026-4308 vLLM: Harmony tool continuations drop `cache_salt` — restoring a cross-tenant prefix-cache membership oracle

Affected - Ecosystem / package: pip / vllm - Affected versions: vLLM ≤ 0.25.1 confirmed on 0.25.1, commit 752a3a504485. The lower bound predates 0.25.1; maintainers can confirm how far back the tool-continuation re-submission has omitted the salt. Summary On the GPT-OSS "Harmony" path POST...

GHSA-935W-9G4M-P28P vLLM: Harmony tool continuations drop `cache_salt` — restoring a cross-tenant prefix-cache membership oracle

Affected - Ecosystem / package: pip / vllm - Affected versions: vLLM ≤ 0.25.1 confirmed on 0.25.1, commit 752a3a504485. The lower bound predates 0.25.1; maintainers can confirm how far back the tool-continuation re-submission has omitted the salt. Summary On the GPT-OSS "Harmony" path POST...

vLLM: Harmony tool continuations drop `cache_salt` — restoring a cross-tenant prefix-cache membership oracle

Affected - Ecosystem / package: pip / vllm - Affected versions: vLLM ≤ 0.25.1 confirmed on 0.25.1, commit 752a3a504485. The lower bound predates 0.25.1; maintainers can confirm how far back the tool-continuation re-submission has omitted the salt. Summary On the GPT-OSS "Harmony" path POST...

CVE-2026-105752 vLLM: Harmony tool continuations drop `cache_salt` — restoring a cross-tenant prefix-cache membership oracle

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 cachesalt value, placing the continuation prefix in the global unsalted cache...

CVE-2026-105752 vLLM: Harmony tool continuations drop `cache_salt` — restoring a cross-tenant prefix-cache membership oracle

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 cachesalt value, placing the continuation prefix in the global unsalted cache...

CVE-2026-105752

The vLLM inference and serving engine is vulnerable in versions prior to 0.30.0 due to a flaw in how Harmony tool continuations are handled via POST /v1/responses requests. The application fails to preserve the cache_salt value when rebuilding the next-turn engine input, causing the continuation ...

CVE-2026-105752 vLLM: Harmony tool continuations drop `cache_salt` — restoring a cross-tenant prefix-cache membership oracle

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 cachesalt value, placing the continuation prefix in the global unsalted cache...

Allocation of Resources Without Limits or Throttling

Overview vllm is an A high-throughput and memory-efficient inference and serving engine for LLMs Affected versions of this package are vulnerable to Allocation of Resources Without Limits or Throttling via the cachesalt field across multiple API protocol handlers Anthropic, OpenAI chat completion...

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

...

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

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-i...

vLLM uses Python 3.12 built-in hash() which leads to predictable hash collisions in prefix cache

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 hashNon...