vLLM versions >= 0.10.2 and < 0.13.0 are missing sparse tensor validation in multimodal embeddings processing. Because PyTorch disables sparse tensor invariant checks by default, an attacker can submit crafted embedding requests with malformed (negative or out-of-bounds) tensor indices, when the prompt-embeds feature is enabled, to trigger crashes or resource exhaustion (denial of service), with potential for out-of-bounds/write-what-where memory corruption. This continues CVE-2025-62164, whose prior fix only disabled the feature by default rather than addressing the root cause.

vLLM is an inference and serving engine for large language models (LLMs). From 0.1.0 to before 0.10.1.1, a Denial of Service (DoS) vulnerability can be triggered by sending a single HTTP GET request with an extremely large header to an HTTP endpoint. This results in server memory exhaustion, potentially leading to a crash or unresponsiveness. The attack does not require authentication, making it exploitable by any remote user. This vulnerability is fixed in 0.10.1.1.

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Versions starting from 0.5.2 and prior to 0.8.5 are vulnerable to denial of service and data exposure via ZeroMQ on multi-node vLLM deployment. In a multi-node vLLM deployment, vLLM uses ZeroMQ for some multi-node communication purposes. The primary vLLM host opens an XPUB ZeroMQ socket and binds it to ALL interfaces. While the socket is always opened for a multi-node deployment, it is only used when doing tensor parallelism across multiple hosts. Any client with network access to this host can connect to this XPUB socket unless its port is blocked by a firewall. Once connected, these arbitrary clients will receive all of the same data broadcasted to all of the secondary vLLM hosts. This data is internal vLLM state information that is not useful to an attacker. By potentially connecting to this socket many times and not reading data published to them, an attacker can also cause a denial of service by slowing down or potentially blocking the publisher. This issue has been patched in version 0.8.5.

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Prior to 0.24.0, the structured_outputs.regex API parameter passes a user-supplied regular expression string directly to the grammar compiler backends with no compilation timeout; in the xgrammar backend the string reaches the regex compiler with no guard, and in the outlines backend the validation step blocks structural issues such as lookarounds and backreferences but performs no complexity analysis, so a pattern with nested quantifiers passes all checks and causes exponential state-space expansion, allowing a single request containing an adversarial regex to hang an inference worker indefinitely and deny service. This issue is fixed in version 0.24.0.

vLLM versions 0.8.0 and later are vulnerable to an Out-of-Memory (OOM) Denial of Service (DoS) attack due to unbounded frame count processing in the `VideoMediaIO.load_base64()` method. When processing `video/jpeg` data URLs, the method splits the base64 data string on commas to extract individual JPEG frames without enforcing a frame count limit. An attacker can exploit this by crafting a single API request containing thousands of comma-separated base64-encoded JPEG frames in a data URL, causing the server to decode all frames into memory and crash due to excessive memory consumption. This vulnerability is reachable via the OpenAI-compatible chat completions API and does not require authentication.

vLLM is an inference and serving engine for large language models (LLMs). In versions from 0.6.4 to before 0.12.0, users can crash the vLLM engine serving multimodal models that use the Idefics3 vision model implementation by sending a specially crafted 1x1 pixel image. This causes a tensor dimension mismatch that results in an unhandled runtime error, leading to complete server termination. This issue has been patched in version 0.12.0.

vLLM versions >= 0.6.3 and < 0.9.0 contain multiple regular expression denial of service (ReDoS) vulnerabilities. Several regex patterns — in vllm/lora/utils.py, the phi4mini tool parser, and the OpenAI-compatible serving chat endpoint — are susceptible to catastrophic backtracking. An attacker submitting crafted input with nested or repeated structures can trigger severe CPU consumption and performance degradation, resulting in denial of service.

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Versions starting from 0.8.0 and prior to 0.8.5 are affected by a critical performance vulnerability in the input preprocessing logic of the multimodal tokenizer. The code dynamically replaces placeholder tokens (e.g., <|audio_|>, <|image_|>) with repeated tokens based on precomputed lengths. Due to ​​inefficient list concatenation operations​​, the algorithm exhibits ​​quadratic time complexity (O(n²))​​, allowing malicious actors to trigger resource exhaustion via specially crafted inputs. This issue has been patched in version 0.8.5.

vLLM is an inference and serving engine for large language models (LLMs). In version 0.8.0 up to but excluding 0.9.0, the vLLM backend used with the /v1/chat/completions OpenAPI endpoint fails to validate unexpected or malformed input in the "pattern" and "type" fields when the tools functionality is invoked. These inputs are not validated before being compiled or parsed, causing a crash of the inference worker with a single request. The worker will remain down until it is restarted. Version 0.9.0 fixes the issue.

vLLM is an inference and serving engine for large language models (LLMs). From 0.1.0 to before 0.19.0, a Denial of Service vulnerability exists in the vLLM OpenAI-compatible API server. Due to the lack of an upper bound validation on the n parameter in the ChatCompletionRequest and CompletionRequest Pydantic models, an unauthenticated attacker can send a single HTTP request with an astronomically large n value. This completely blocks the Python asyncio event loop and causes immediate Out-Of-Memory crashes by allocating millions of request object copies in the heap before the request even reaches the scheduling queue. This vulnerability is fixed in 0.19.0.

vLLM is an inference and serving engine for large language models (LLMs). From version 0.5.5 to before version 0.18.0, Librosa defaults to using numpy.mean for mono downmixing (to_mono), while the international standard ITU-R BS.775-4 specifies a weighted downmixing algorithm. This discrepancy results in inconsistency between audio heard by humans (e.g., through headphones/regular speakers) and audio processed by AI models (Which infra via Librosa, such as vllm, transformer). This issue has been patched in version 0.18.0.

vLLM is an inference and serving engine for large language models (LLMs). From versions 0.10.2 to before 0.11.1, a memory corruption vulnerability could lead to a crash (denial-of-service) and potentially remote code execution (RCE), exists in the Completions API endpoint. When processing user-supplied prompt embeddings, the endpoint loads serialized tensors using torch.load() without sufficient validation. Due to a change introduced in PyTorch 2.8.0, sparse tensor integrity checks are disabled by default. As a result, maliciously crafted tensors can bypass internal bounds checks and trigger an out-of-bounds memory write during the call to to_dense(). This memory corruption can crash vLLM and potentially lead to code execution on the server hosting vLLM. This issue has been patched in version 0.11.1.

The read_mru_list function in NTP before 4.2.8p9 allows remote attackers to cause a denial of service (crash) via a crafted mrulist query.

An Improper Validation of Syntactic Correctness of Input vulnerability in the kernel of Juniper Networks Junos OS Evolved on PTX series allows a network-based, unauthenticated attacker to cause a Denial of Service (DoS). When an incoming TCP packet destined to the device is malformed there is a possibility of a kernel panic. Only TCP packets destined to the ports for BGP, LDP and MSDP can trigger this. This issue only affects PTX10004, PTX10008, PTX10016. No other PTX Series devices or other platforms are affected. This issue affects Juniper Networks Junos OS Evolved: 20.4-EVO versions prior to 20.4R3-S4-EVO; 21.3-EVO versions prior to 21.3R3-EVO; 21.4-EVO versions prior to 21.4R3-EVO; 22.1-EVO versions prior to 22.1R2-EVO. This issue does not affect Juniper Networks Junos OS Evolved versions prior to 20.4R1-EVO.

Improper input validation in the DAL subsystem for Intel(R) CSME versions before 12.0.64, 13.0.32, 14.0.33 and 14.5.12 may allow an unauthenticated user to potentially enable denial of service via network access.

In Message and toBundle of Notification.java, there is a possible UI slowdown or crash due to improper input validation. This could lead to remote denial of service if a malicious contact file is received, with no additional execution privileges needed. User interaction is not needed for exploitation.Product: AndroidVersions: Android-10 Android-11 Android-8.0 Android-8.1 Android-9Android ID: A-147358092

An issue was discovered in Samsung Mobile Processor, Wearable Processor, and Modem (Exynos 980, 850, 990, 1080, 2100, 1280, 2200, 1330, 1380, 1480, 2400, 1580, 2500, 1680, 9110, W920, W930, W1000, Modem 5123, Modem 5300, Modem 5400, and Modem 5410). The absence of proper input validation leads to a Denial of Service.

In Tensorflow before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, changing the TensorFlow's `SavedModel` protocol buffer and altering the name of required keys results in segfaults and data corruption while loading the model. This can cause a denial of service in products using `tensorflow-serving` or other inference-as-a-service installments. Fixed were added in commits f760f88b4267d981e13f4b302c437ae800445968 and fcfef195637c6e365577829c4d67681695956e7d (both going into TensorFlow 2.2.0 and 2.3.0 but not yet backported to earlier versions). However, this was not enough, as #41097 reports a different failure mode. The issue is patched in commit adf095206f25471e864a8e63a0f1caef53a0e3a6, and is released in TensorFlow versions 1.15.4, 2.0.3, 2.1.2, 2.2.1, or 2.3.1.

Brocade SANnav before v.2.1.0a could allow remote attackers cause a denial-of-service condition due to a lack of proper validation, of the length of user-supplied data as name for custom field name.

An Improper Validation of Specified Type of Input vulnerability in the routing protocol daemon (rpd) of Juniper Networks Junos OS allows an attacker to cause an RPD memory leak leading to a Denial of Service (DoS). This memory leak only occurs when the attacker's packets are destined to any configured IPv6 address on the device. This issue affects: Juniper Networks Junos OS 21.1 versions prior to 21.1R3-S2; 21.2 versions prior to 21.2R3-S1; 21.3 versions prior to 21.3R3; 21.4 versions prior to 21.4R2; 22.1 versions prior to 22.1R2. This issue does not affect Juniper Networks Junos OS versions prior to 21.1R1.

HCL Domino is susceptible to a Denial of Service (DoS) vulnerability due to insufficient validation of input to its public API. An unauthenticated attacker could could exploit this vulnerability to crash the Domino server.

An issue was discovered in RRC in Samsung Mobile Processor, Wearable Processor, and Modem Exynos 980, 990, 850, 1080, 2100, 1280, 2200, 1330, 1380, 1480, 2400, 1580, 2500, 9110, W920, W930, W1000, Modem 5123, Modem 5300, and Modem 5400. Improper memory initialization results in an illegal memory access, causing a system crash via a malformed RRCReconfiguration message.

In bindArtworkAndColors of MediaControlPanel.java, there is a possible way to crash the phone due to improper input validation. This could lead to remote denial of service with no additional execution privileges needed. User interaction is not needed for exploitation.Product: AndroidVersions: Android-13Android ID: A-239368697

A denial-of-service vulnerability exists in the Rockwell Automation affected products when specially crafted packets are sent to the CIP Security Object. If exploited the device will become unavailable and require a factory reset to recover.

HCL Domino is susceptible to a Denial of Service vulnerability due to improper validation of user-supplied input, potentially giving an attacker the ability to crash the server. Versions previous to release 9.0.1 FP10 IF6 and release 10.0.1 are affected.

HCL Domino is susceptible to a Denial of Service vulnerability caused by improper validation of user-supplied input. A remote unauthenticated attacker could exploit this vulnerability using a specially-crafted email message to hang the server. Versions previous to releases 9.0.1 FP10 IF6, 10.0.1 FP5 and 11.0.1 are affected.

NaviServer 4.99.4 to 4.99.19 allows denial of service due to the nsd/driver.c ChunkedDecode function not properly validating the length of a chunk. A remote attacker can craft a chunked-transfer request that will result in a negative value being passed to memmove via the size parameter, causing the process to crash.

CServer::SendMsg in engine/server/server.cpp in Teeworlds 0.7.x before 0.7.5 allows remote attackers to shut down the server.

A Denial of Service vulnerability has been identified in FlexNet Publisher's lmadmin.exe version 11.16.6. A certain message protocol can be exploited to cause lmadmin to crash.

In modem, there is a possible system crash due to improper input validation. This could lead to remote escalation of privilege with no additional execution privileges needed.

This vulnerability allows remote attackers to create a denial-of-service condition on affected installations of C-MORE HMI EA9 Firmware version 6.52 touch screen panels. Authentication is not required to exploit this vulnerability. The specific flaw exists within the EA-HTTP.exe process. The issue results from the lack of proper input validation prior to further processing user requests. An attacker can leverage this vulnerability to create a denial-of-service condition on the system. Was ZDI-CAN-10527.

Buffer over-read while parsing RPS due to lack of check of input validation on values received from user side. in Snapdragon Auto, Snapdragon Compute, Snapdragon Connectivity, Snapdragon Consumer IOT, Snapdragon Industrial IOT, Snapdragon Mobile

The HTTP server in Mongoose before 7.10 accepts requests containing negative Content-Length headers. By sending a single attack payload over TCP, an attacker can cause an infinite loop in which the server continuously reparses that payload, and does not respond to any other requests.

A denial of service issue was addressed with improved input validation. This issue is fixed in iOS 13.5 and iPadOS 13.5, macOS Catalina 10.15.5. A remote attacker may be able to cause a denial of service.

OpenBao exists to provide a software solution to manage, store, and distribute sensitive data including secrets, certificates, and keys. OpenBao before v2.3.0 allowed an attacker to perform unauthenticated, unaudited cancellation of root rekey and recovery rekey operations, effecting a denial of service. In OpenBao v2.2.0 and later, manually setting the configuration option `disable_unauthed_rekey_endpoints=true` allows an operator to deny these rarely-used endpoints on global listeners. A patch is available at commit fe75468822a22a88318c6079425357a02ae5b77b. In a future OpenBao release communicated on OpenBao's website, the maintainers will set this to `true` for all users and provide an authenticated alternative. As a workaround, if an active proxy or load balancer sits in front of OpenBao, an operator can deny requests to these endpoints from unauthorized IP ranges.

Transient DOS while processing SMS container of non-standard size received in DL NAS transport in NR.

A Denial of Service (Dos) vulnerability in Nozomi Networks Guardian and CMC, due to improper input validation in certain fields used in the Asset Intelligence functionality of our IDS, allows an unauthenticated attacker to crash the IDS module by sending specially crafted malformed network packets. During the (limited) time window before the IDS module is automatically restarted, network traffic may not be analyzed.

Transient DOS while processing DL NAS Transport message when message ID is not defined in the 3GPP specification.

Transient DOS while processing PDU Release command with a parameter PDU ID out of range.

parse-server-push-adapter is the official Push Notification adapter for Parse Server. The Parse Server Push Adapter can crash Parse Server due to an invalid push notification payload. This issue has been patched in version 4.1.3.

Improper input validation in subsystem for Intel(R) AMT versions before 11.8.77, 11.12.77, 11.22.77 and 12.0.64 may allow an unauthenticated user to potentially enable denial of service via network access.

Transient DOS while processing CAG info IE received from NW.

In wlan firmware, there is a possible firmware assertion due to improper input handling. This could lead to remote denial of service with no additional execution privileges needed. User interaction is not needed for exploitation. Patch ID: ALPS07932637; Issue ID: ALPS07932637.