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 a high-throughput and memory-efficient inference and serving engine for LLMs. The outlines library is one of the backends used by vLLM to support structured output (a.k.a. guided decoding). Outlines provides an optional cache for its compiled grammars on the local filesystem. This cache has been on by default in vLLM. Outlines is also available by default through the OpenAI compatible API server. The affected code in vLLM is vllm/model_executor/guided_decoding/outlines_logits_processors.py, which unconditionally uses the cache from outlines. A malicious user can send a stream of very short decoding requests with unique schemas, resulting in an addition to the cache for each request. This can result in a Denial of Service if the filesystem runs out of space. Note that even if vLLM was configured to use a different backend by default, it is still possible to choose outlines on a per-request basis using the guided_decoding_backend key of the extra_body field of the request. This issue applies only to the V0 engine and is fixed in 0.8.0.

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 a library for LLM inference and serving. From 0.12.0 to before 0.24.0, sending a pure prompt embeds payload in a /v1/completions request with a model using M-RoPE causes EngineCore to fail an assertion and fatally crash, shutting down the entire server application. Any remote user who is authorized to make a /v1/completions request can make such a request and induce a crash. This issue is fixed in version 0.24.0.

vLLM is an inference and serving engine for large language models. From 0.22.0 to 0.23.0, the /v1/audio/transcriptions and /v1/audio/translations routes call request.file.read() to fully materialize an uploaded audio file into memory before vLLM checks the documented VLLM_MAX_AUDIO_CLIP_FILESIZE_MB compressed upload size limit (default 25 MB) later in the speech-to-text preprocessing step, so an API caller who can reach those routes can submit an oversized multipart upload and cause vLLM to allocate memory proportional to the uploaded file size before the request is rejected as too large, creating memory pressure or terminating the process depending on deployment resource limits. This issue is fixed in version 0.24.0.

vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.23.1rc0, vLLM's /v1/audio/transcriptions endpoint limits compressed upload size but not decoded PCM output. A 25MB OPUS file expands to ~14.9GB of float32 PCM at decode time. This vulnerability is fixed in 0.23.1rc0.

vLLM is an inference and serving engine for large language models (LLMs). From 0.18.0 to before 0.20.0, the extract_hidden_states speculative decoding proposer in vLLM returns a tensor with an incorrect shape after the first decode step, causing a RuntimeError that crashes the EngineCore process. The crash is triggered when any request in the batch uses sampling penalty parameters (repetition_penalty, frequency_penalty, or presence_penalty). A single request with a penalty parameter (e.g., "repetition_penalty": 1.1) is sufficient to crash the server. This vulnerability is fixed in 0.20.0.

vLLM is an inference and serving engine for large language models (LLMs). From 0.6.1 to before 0.20.0, there is a a Token Injection vulnerability in vLLM’s multimodal processing. Unauthenticated, text-only prompts that spell special tokens are interpreted as control. Image and video placeholder sequences supplied without matching data cause vLLM to index into empty grids during input-position computation, raising an unhandled IndexError and terminating the worker or degrading availability. Multimodal paths that rely on image_grid_thw/video_grid_thw are affected. This vulnerability is fixed in 0.20.0.

vLLM is an inference and serving engine for large language models (LLMs). From 0.7.0 to before 0.19.0, the VideoMediaIO.load_base64() method at vllm/multimodal/media/video.py splits video/jpeg data URLs by comma to extract individual JPEG frames, but does not enforce a frame count limit. The num_frames parameter (default: 32), which is enforced by the load_bytes() code path, is completely bypassed in the video/jpeg base64 path. An attacker can send a single API request containing thousands of comma-separated base64-encoded JPEG frames, causing the server to decode all frames into memory and crash with OOM. This vulnerability is fixed in 0.19.0.

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). 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 is an inference and serving engine for large language models (LLMs). Version 0.8.0 up to but excluding 0.9.0 have a Denial of Service (ReDoS) that causes the vLLM server to crash if an invalid regex was provided while using structured output. This vulnerability is similar to GHSA-6qc9-v4r8-22xg/CVE-2025-48942, but for regex instead of a JSON schema. Version 0.9.0 fixes the issue.

vLLM is an inference and serving engine for large language models (LLMs). From version 0.5.5 to before 0.11.1, users can crash the vLLM engine serving multimodal models by passing multimodal embedding inputs with correct ndim but incorrect shape (e.g. hidden dimension is wrong), regardless of whether the model is intended to support such inputs (as defined in the Supported Models page). This issue has been patched in version 0.11.1.

vLLM is an inference and serving engine for large language models (LLMs). From version 0.5.5 to before 0.11.1, the /v1/chat/completions and /tokenize endpoints allow a chat_template_kwargs request parameter that is used in the code before it is properly validated against the chat template. With the right chat_template_kwargs parameters, it is possible to block processing of the API server for long periods of time, delaying all other requests. This issue has been patched in version 0.11.1.

vLLM is an inference and serving engine for large language models (LLMs). In versions 0.8.0 up to but excluding 0.9.0, hitting the /v1/completions API with a invalid json_schema as a Guided Param kills the vllm server. This vulnerability is similar GHSA-9hcf-v7m4-6m2j/CVE-2025-48943, but for regex instead of a JSON schema. Version 0.9.0 fixes the issue.

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 >= 0.6.3 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.

A vulnerability was found in the Keycloak-services package. If untrusted data is passed to the SearchQueryUtils method, it could lead to a denial of service (DoS) scenario by exhausting system resources due to a Regex complexity.

An issue has been discovered in GitLab CE/EE affecting all versions starting from 16.9 prior to 16.9.7, starting from 16.10 prior to 16.10.5, and starting from 16.11 prior to 16.11.2. A problem with the processing logic for Discord Integrations Chat Messages can lead to a regular expression DoS attack on the server.

An issue has been discovered in GitLab affecting all versions starting from 13.9 before 14.8.6, all versions starting from 14.9 before 14.9.4, all versions starting from 14.10 before 14.10.1. GitLab was not correctly handling malicious text in the CI Editor and CI Pipeline details page allowing the attacker to cause uncontrolled resource consumption.

Zohocorp ManageEngine Exchange Reporter Plus through 5721 are vulnerable to ReDOS vulnerability in the search module.

Pattern Redirects in Liferay Portal 7.4.3.48 through 7.4.3.76, and Liferay DXP 7.4 update 48 through 76 allows regular expressions that are vulnerable to ReDoS attacks to be used as patterns, which allows remote attackers to consume an excessive amount of server resources via crafted request URLs.

An issue has been discovered in GitLab EE affecting all versions before 16.8.6, all versions starting from 16.9 before 16.9.4, all versions starting from 16.10 before 16.10.2. It was possible for an attacker to cause a denial of service using malicious crafted content in a junit test report file.

A denial of service vulnerability was identified in GitLab CE/EE, versions 16.7.7 prior to 16.8.6, 16.9 prior to 16.9.4 and 16.10 prior to 16.10.2 which allows an attacker to spike the GitLab instance resources usage resulting in service degradation via chat integration feature.

An issue has been discovered in GitLab affecting all versions starting from 15.11 before 16.1.5, all versions starting from 16.2 before 16.2.5, all versions starting from 16.3 before 16.3.1. An authenticated user could trigger a denial of service when importing or cloning malicious content.

An issue has been discovered in GitLab affecting all versions starting from 15.11 before 16.1.5, all versions starting from 16.2 before 16.2.5, all versions starting from 16.3 before 16.3.1. An authenticated user could trigger a denial of service when importing or cloning malicious content.

An issue has been discovered in GitLab CE/EE affecting all versions from 12.7 prior to 16.6.6, 16.7 prior to 16.7.4, and 16.8 prior to 16.8.1 It was possible for an attacker to trigger a Regular Expression Denial of Service via a `Cargo.toml` containing maliciously crafted input.

formula is a math and string formula parser. In versions prior to 3.0.1 crafted user-provided strings to formula's parser might lead to polynomial execution time and a denial of service. Users should upgrade to 3.0.1+. There are no known workarounds for this vulnerability.

Discourse is an open source discussion platform. In affected versions a malicious user can cause a regular expression denial of service using a carefully crafted git URL. This issue is patched in the latest stable, beta and tests-passed versions of Discourse. Users are advised to upgrade. There are no known workarounds for this issue.

An issue has been discovered in GitLab affecting all versions starting from 15.10 before 16.1, leading to a ReDoS vulnerability in the Jira prefix

Zulip is an open source team chat server. In affected versions Zulip allows organization administrators on a server to configure "linkifiers" that automatically create links from messages that users send, detected via arbitrary regular expressions. Malicious organization administrators could subject the server to a denial-of-service via regular expression complexity attacks; most simply, by configuring a quadratic-time regular expression in a linkifier, and sending messages that exploited it. A regular expression attempted to parse the user-provided regexes to verify that they were safe from ReDoS -- this was both insufficient, as well as _itself_ subject to ReDoS if the organization administrator entered a sufficiently complex invalid regex. Affected users should [upgrade to the just-released Zulip 4.7](https://zulip.readthedocs.io/en/latest/production/upgrade-or-modify.html#upgrading-to-a-release), or [`main`](https://zulip.readthedocs.io/en/latest/production/upgrade-or-modify.html#upgrading-from-a-git-repository).

Self-ReDoS (Regular expression Denial of Service) exists with Role Name search field of Kaleo Designer portlet JavaScript in Liferay Portal 7.4.0 through 7.4.3.131, and Liferay DXP 2024.Q4.0 through 2024.Q4.1, 2024.Q3.0 through 2024.Q3.13, 2024.Q2.1 through 2024.Q2.13, 2024.Q1.1 through 2024.Q1.20 and 7.4 GA through update 92, which allows authenticated users with permissions to update Kaleo Workflows to enter a malicious Regex pattern causing their browser to hang for a very long time.

Mattermost versions 11.7.x <= 11.7.2, 11.6.x <= 11.6.4, 10.11.x <= 10.11.19 fail to validate the length and content of message attachment field values, which allows an authenticated attacker to cause a denial of service for all users in a channel via a post containing a specially crafted payload that triggers catastrophic backtracking in the client-side markdown parser.. Mattermost Advisory ID: MMSA-2026-00658

Open WebUI is an extensible, feature-rich, and user-friendly self-hosted AI platform. From 0.9.2 before 0.10.0, the SKILL_MENTION_RE and strip_re regular expressions in backend/open_webui/utils/middleware.py parsed <$skillId|label> skill mentions with overlapping quantifiers, allowing an authenticated chat message containing <$ without a closing > to trigger quadratic backtracking and block the asyncio event loop. This issue is fixed in version 0.10.0.

LobeChat before version 2.2.10-canary.15 contains a regular expression denial of service (ReDoS) vulnerability that allows authenticated attackers to block the Node.js event loop by supplying a catastrophic-backtracking pattern in a GitHub repository URL path during skill import. Attackers can craft a malicious basePath value containing unescaped regex metacharacters such as catastrophic-backtracking patterns, which are injected into a dynamically constructed regular expression in the findSkillMd function and executed synchronously against archive entries, denying service to all concurrent users for tens of seconds per request.

Grav before 2.0.4 contains a regular expression denial of service (ReDoS) vulnerability in the regex_replace filter and function, which are allowlisted in the Twig content sandbox. When Twig processing in page content is enabled (security.twig_content.process_enabled: true, disabled by default), an authenticated page editor can supply a catastrophically backtracking PCRE pattern that is passed directly to PHP's preg_replace(), causing unbounded CPU consumption and denial of service to the web server process.

An issue has been discovered in GitLab CE/EE affecting all versions from 13.2 before 18.0.6, 18.1 before 18.1.4, and 18.2 before 18.2.2 that could have allowed authenticated users to create a denial of service condition by sending specially crafted markdown payloads to the Wiki feature.

An issue has been discovered in GitLab CE/EE affecting all versions starting from 12.3 before 16.3.6, all versions starting from 16.4 before 16.4.2, all versions starting from 16.5 before 16.5.1. A Regular Expression Denial of Service was possible by adding a large string in timeout input in gitlab-ci.yml file.

Nautobot is a Network Source of Truth and Network Automation Platform. Prior to 2.4.33 and 3.1.2, Nautobot UI object-bulk-rename endpoints (for example, /dcim/interfaces/rename/) were vulnerable to application-wide denial of service via maliciously crafted regular expressions in the find field in combination with the use_regex flag. This vulnerability is fixed in 2.4.33 and 3.1.2.

In versions 3.1.0 and lower of the Splunk Supporting Add-on for Active Directory, also known as SA-ldapsearch, a vulnerable regular expression pattern could lead to a Regular Expression Denial of Service (ReDoS) attack.

A vulnerability classified as problematic was found in Langflow up to 1.0.18. Affected by this vulnerability is an unknown functionality of the file \src\backend\base\langflow\interface\utils.py of the component HTTP POST Request Handler. The manipulation of the argument remaining_text leads to inefficient regular expression complexity. The exploit has been disclosed to the public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.

In JetBrains YouTrack before 2024.3.52635 potential ReDoS was possible due to vulnerable RegExp in Ruby syntax detector

A Denial of Service (DoS) condition has been discovered in GitLab CE/EE affecting all versions from 7.10 prior before 16.11.5, version 17.0 before 17.0.3, and 17.1 before 17.1.1. It is possible for an attacker to cause a denial of service using a crafted markdown page.

Rack is a modular Ruby web server interface. Starting in version 3.1.0 and prior to version 3.1.5, Regular Expression Denial of Service (ReDoS) vulnerability exists in the `Rack::Request::Helpers` module when parsing HTTP Accept headers. This vulnerability can be exploited by an attacker sending specially crafted `Accept-Encoding` or `Accept-Language` headers, causing the server to spend excessive time processing the request and leading to a Denial of Service (DoS). The fix for CVE-2024-26146 was not applied to the main branch and thus while the issue was fixed for the Rack v3.0 release series, it was not fixed in the v3.1 release series until v3.1.5. Users of versions on the 3.1 branch should upgrade to version 3.1.5 to receive the fix.

Wagtail is an open source content management system built on Django. A bug in Wagtail's `parse_query_string` would result in it taking a long time to process suitably crafted inputs. When used to parse sufficiently long strings of characters without a space, `parse_query_string` would take an unexpectedly large amount of time to process, resulting in a denial of service. In an initial Wagtail installation, the vulnerability can be exploited by any Wagtail admin user. It cannot be exploited by end users. If your Wagtail site has a custom search implementation which uses `parse_query_string`, it may be exploitable by other users (e.g. unauthenticated users). Patched versions have been released as Wagtail 5.2.6, 6.0.6 and 6.1.3.

uri-js is a module that tries to fully implement RFC 3986. One of these features is validating whether or not a supplied URL is valid or not. To do this, uri-js uses a regular expression, This regular expression is vulnerable to redos. This causes the program to hang and the CPU to idle at 100% usage while uri-js is trying to validate if the supplied URL is valid or not. To check if you're vulnerable, look for a call to `require("uri-js").parse()` where a user is able to send their own input. This affects uri-js 2.1.1 and earlier.

fast-jwt provides fast JSON Web Token (JWT) implementation. From 5.0.0 to 6.2.0, a denial-of-service condition exists in fast-jwt when the allowedAud verification option is configured using a regular expression. Because the aud claim is attacker-controlled and the library evaluates it against the supplied RegExp, a crafted JWT can trigger catastrophic backtracking in the JavaScript regex engine, resulting in significant CPU consumption during verification. This vulnerability is fixed in 6.2.1.

PraisonAI is a multi-agent teams system. Prior to version 4.5.90, MCPToolIndex.search_tools() compiles a caller-supplied string directly as a Python regular expression with no validation, sanitization, or timeout. A crafted regex causes catastrophic backtracking in the re engine, blocking the Python thread for hundreds of seconds and causing a complete service outage. This issue has been patched in version 4.5.90.

ReDoS flaw in RefMatcher when matching branch names using wildcards in GitLab EE/CE affecting all versions from 11.3 prior to 17.0.6, 17.1 prior to 17.1.4, and 17.2 prior to 17.2.2 allows denial of service via Regex backtracking.

An issue has been discovered in GitLab CE/EE affecting all versions before 16.9.7, all versions starting from 16.10 before 16.10.5, all versions starting from 16.11 before 16.11.2. It was possible for an attacker to cause a denial of service using maliciously crafted markdown content.