CVE-2025-62426 vLLM vulnerable to DoS via large Chat Completion or Tokenization requests with specially crafted `chat_template_kwargs`

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 chattemplatekwargs request parameter that is used in the code before it is properly validated against the chat template. With the...

EUVD-2025-198356

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 chattemplatekwargs request parameter that is used in the code before it is properly validated against the chat template. With the...

CVE-2025-62426

Summary: CVE-2025-62426 affects vLLM up to versions before 0.11.1. The /v1/chat/completions and /tokenize endpoints accept a chat_template_kwargs parameter that is used before validation, allowing an attacker to block the API server by forcing large tokenization tasks and delaying all other reque...

CVE-2025-62426 vLLM vulnerable to DoS via large Chat Completion or Tokenization requests with specially crafted `chat_template_kwargs`

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 chattemplatekwargs request parameter that is used in the code before it is properly validated against the chat template. With the...

GHSA-69J4-GRXJ-J64P vLLM vulnerable to DoS via large Chat Completion or Tokenization requests with specially crafted `chat_template_kwargs`

Summary The /v1/chat/completions and /tokenize endpoints allow a chattemplatekwargs request parameter that is used in the code before it is properly validated against the chat template. With the right chattemplatekwargs parameters, it is possible to block processing of the API server for long...

TencentOS Server 4: grafana (TSSA-2025:0339)

The version of Tencent Linux installed on the remote TencentOS Server 4 host is prior to tested version. It is, therefore, affected by multiple vulnerabilities as referenced in the TSSA-2025:0339 advisory. Package updates are available for TencentOS Server 4 that fix the following vulnerabilities...

OESA-2025-2373 llama.cpp security update

Security Fixes: llama.cpp is an inference of several LLM models in C/C++. Prior to version b5721, there is a signed vs. unsigned integer overflow in llama.cpp's tokenizer implementation llamavocab::tokenize src/llama-vocab.cpp:3036 resulting in unintended behavior in tokens copying size compariso...

EUVD-2021-0148

Malware in sbrugna...

Imperceptible Jailbreaking against Large Language Models

Jailbreaking attacks on the vision modality typically rely on imperceptible adversarial perturbations, whereas attacks on the textual modality are generally assumed to require visible modifications e.g., non-semantic suffixes. In this paper, we introduce imperceptible jailbreaks that exploit a...

EUVD-2022-29941

Malicious code in bioql PyPI...

text-generation-inference: Unbounded external image fetch in validation leads to resource-exhaustion DoS

Description Text Generation Inference Router DoS via pre-validation image fetch in VLM mode. Affected: Router workspace version 3.3.6 the latest repo, when deployed with a vision/VLM model e.g., Idefics/Mllama/Idefics2/Idefics3/Gemma3/Llama4/Paligemma/LlavaNext/Qwen2VL/Qwen25VL. Pure text LLMs do...

BERTector: Intrusion Detection Based on Joint-Dataset Learning

Intrusion detection systems IDS are facing challenges in generalization and robustness due to the heterogeneity of network traffic and the diversity of attack patterns. To address this issue, we propose a new joint-dataset training paradigm for IDS and propose a scalable BERTector framework based...

DEBIAN-CVE-2025-52566

llama.cpp is an inference of several LLM models in C/C++. Prior to version b5721, there is a signed vs. unsigned integer overflow in llama.cpp's tokenizer implementation llamavocab::tokenize src/llama-vocab.cpp:3036 resulting in unintended behavior in tokens copying size comparison. Allowing...

Out-of-Bounds

Overview Affected versions of this package are vulnerable to Out-of-Bounds via the tokenize function in the file llama-vocab.cpp. An attacker can achieve arbitrary code execution or cause a denial of service by providing specially crafted text input that triggers a heap overflow during the...

CVE-2025-52566

CVE-2025-52566 affects llama.cpp (tokenizer in llama_vocab::tokenize). Prior to version b5721, a signed vs. unsigned integer overflow enables heap-overflow during tokenization, causing unintended behavior in token copying. This could enable heap-related issues during inference with crafted input....

PT-2025-26690 · Llama.Cpp · Llama.Cpp

Name of the Vulnerable Software and Affected Versions: llama.cpp version prior to b5721 Description: The issue is related to a signed vs. unsigned integer overflow in the llama vocab::tokenize function of llama.cpp's tokenizer implementation, resulting in unintended behavior during token copying...

Medium: amazon-cloudwatch-agent

Issue Overview: The tokenizer incorrectly interprets tags with unquoted attribute values that end with a solidus character / as self-closing. When directly using Tokenizer, this can result in such tags incorrectly being marked as self-closing, and when using the Parse functions, this can result i...

Medium: amazon-cloudwatch-agent

Issue Overview: The tokenizer incorrectly interprets tags with unquoted attribute values that end with a solidus character / as self-closing. When directly using Tokenizer, this can result in such tags incorrectly being marked as self-closing, and when using the Parse functions, this can result i...

New TokenBreak Attack Bypasses AI Moderation with Single-Character Text Changes

Cybersecurity researchers have discovered a novel attack technique called TokenBreak that can be used to bypass a large language model's LLM safety and content moderation guardrails with just a single character change. "The TokenBreak attack targets a text classification model's tokenization...

Joint-GCG: Unified Gradient-Based Poisoning Attacks on Retrieval-Augmented Generation Systems

Retrieval-Augmented Generation RAG systems enhance Large Language Models LLMs by retrieving relevant documents from external corpora before generating responses. This approach significantly expands LLM capabilities by leveraging vast, up-to-date external knowledge. However, this reliance on...