Cache Poisoning

vLLM is vulnerable to Cache Poisoning. The vulnerability is due to hash collisions due to the use of Python's built-in hash function for prefix caching, which makes hashNone a predictable constant value, allowing an attacker to intentionally populate the cache with colliding prompts and interfere...

Description of the security update for Office 2016: February 11, 2025 (KB5002686)

None None...

Description of the security update for Office 2016: February 11, 2025 (KB5002684)

None None...

Description of the security update for SharePoint Server 2019: February 11, 2025 (KB5002678)

None None...

NTLM Hash Disclosure Spoofing Vulnerability

...

Astra Linux – Vulnerability in Linux 6.1

In the Linux kernel, the following vulnerability has been resolved: ima: fixed the buffer overflow in imaeventdigestinitcommon. The function imaeventdigestinit calls imaeventdigestinitcommon, with HASHALGOLAST being used to access the array hashdigestsize. This leads to a buffer overflow. A...

Astra Linux – Vulnerability in Linux 6.1

In the Linux kernel, the following vulnerability has been resolved: net/xen-netback: This issue prevents a Use-after-Free UAF condition in xenvifFlushHash. During the listforeachentryrcu iteration, when calling xenvifFlushHash, kfreercu does not exist within the rcu read critical section...

Astra Linux – Vulnerability in Linux 6.1

In the Linux kernel, the following vulnerabilities have been resolved: ksmbd: The “binding mark” of a reused connection was unset. Steve French reported a null pointer dereference error from the sha256 lib. The cifs.ko module can send session setup requests using a reused connection. If a reused...

Astra Linux – Vulnerability in Linux 6.1

In the Linux kernel, the following vulnerability has been resolved: posix-timers: Ensure that the limit on the timer ID search loop is valid. The posixtimeradd function attempts to allocate a posix timer ID by starting from the cached ID stored after the last successful allocation. This is done b...

Astra Linux – Vulnerability in Linux 6.1

In the Linux kernel, the following vulnerability has been resolved: sched: schcake: Fixed the logic for accounting bulk flows in order to ensure host fairness. In schcake, we keep track of the number of active bulk flows per host. This is done when running in the dst/src host fairness mode, which...

PT-2025-6320 · Microsoft · Windows

Name of the Vulnerable Software and Affected Versions: Windows versions prior to February 2025 patchday Description: A spoofing issue allows attackers to affect the system. This issue is related to the disclosure of NTLM hash, which can be exploited by attackers. The estimated number of potential...

CVE-2025-25183

A flaw was found in the vllm package. Maliciously constructed statements can lead to hash collisions, resulting in cache reuse, which can interfere with subsequent responses and cause unintended behavior. The impact of a collision would be using a cache that was generated using different content...

PYSEC-2025-62

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...

PYSEC-2025-62

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...

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-i...

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...

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...

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 安全漏洞

vLLM is a high throughput and memory efficient inference and service engine for LLM from the vLLM open source. A security vulnerability exists in vLLM that stems from a maliciously constructed statement that could lead to a hash collision, which could lead to cache reuse, which could interfere wi...

Use of Weak Hash

Overview vllm is an A high-throughput and memory-efficient inference and serving engine for LLMs Affected versions of this package are vulnerable to Use of Weak Hash due to the use of a predictable constant value in the Python 3.12 built-in hash function. An attacker can interfere with subsequent...