On July 7, 2026, OSV published 12 new CVEs targeting the HuggingFace transformers library (sources 77-88).
On July 7, 2026, OSV published 12 new CVEs targeting the HuggingFace transformers library (sources 77-88).
confidence score
Strong evidence: 3 independent source classes support this read.
signal brief
On July 7, 2026, OSV published 12 new CVEs targeting the HuggingFace transformers library (sources 77-88). The vulnerabilities include multiple Regular Expression Denial of Service (ReDoS) flaws in components such as DonutProcessor, SETTING_RE, MarianTokenizer, and AdamWeightDecay optimizer, as well as a deserialization of untrusted data vulnerability (CVE-2024-3568) and an improper input validation issue allowing username injection. These CVEs affect users of the widely adopted transformers library, potentially enabling denial of service or remote code execution in production AI pipelines.
Subsequently, on July 11, HuggingFace released transformers 5.13.1 on PyPI (source 76), which is assumed to contain patches for these vulnerabilities, though the release notes were not explicit. The rapid release suggests the team is actively addressing the issues, but the breadth of the CVEs (12 in one batch) indicates a systemic security review gap.
What the sources said:
- OSV advisory 77: "Transformers is vulnerable to ReDoS attack through its DonutProcessor class."
- OSV advisory 78: "Transformers Deserialization of Untrusted Data vulnerability."
- OSV advisory 86: "Transformers's Improper Input Validation vulnerability can be exploited through username injection."
- PyPI release 76: lists transformers 5.13.1 without detailed changelog, but the timing suggests a security fix release.
source data used
“Downloads: 0 Likes: 2 Tags: code, qlora, checkpoint-fabric, work-in-progress, base_model:Qwen/Qwen3.5-4B, base_model:finetune:Qwen/Qwen3.5-4B, license:apache-2.0, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: transformers, safetensors, gpt2, text-generation, generated_from_trainer, sft, trl, base_model:fpadovani/isl-latn-100mb-ppt-shuff-dyck-100mb_seed10, base_model:finetune:fpadovani/isl-latn-100mb-ppt-shuff-dyck-100mb_seed10, text-generation-inference, endpoints_compatible,”
“Downloads: 0 Likes: 0 Tags: transformers, safetensors, wav2vec2, automatic-speech-recognition, generated_from_trainer, base_model:facebook/mms-300m, base_model:finetune:facebook/mms-300m, license:cc-by-nc-4.0, endpoints_compatible, region:us”
“Downloads: 0 Likes: 3 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: safetensors, fusion-embedding-connector, embeddings, multimodal, audio, retrieval, matryoshka, qwen3-vl, adapters, feature-extraction, en, base_model:Qwen/Qwen3-VL-Embedding-2B”
“Downloads: 0 Likes: 0 Tags: transformers, safetensors, gpt_neox, text-generation, generated_from_trainer, trl, ppo, conversational, dataset:stanfordnlp/imdb, arxiv:1909.08593, base_model:EleutherAI/pythia-160m, base_model:finetune:EleutherAI/pythia-160m”
“Downloads: 0 Likes: 0 Tags: transformers, safetensors, gpt_neox, text-generation, generated_from_trainer, trl, ppo, conversational, dataset:stanfordnlp/imdb, arxiv:1909.08593, base_model:EleutherAI/pythia-160m, base_model:finetune:EleutherAI/pythia-160m”
“Downloads: 57 Likes: 0 Tags: transformers, safetensors, whisper, automatic-speech-recognition, generated_from_trainer, endpoints_compatible, region:us”
“Downloads: 0 Likes: 0 Tags: safetensors, ConvLSTMAutoencoder, model_hub_mixin, pytorch_model_hub_mixin, region:us”
“Downloads: 0 Likes: 0 Tags: transformers, safetensors, gpt_neox, text-generation, generated_from_trainer, ppo, trl, conversational, dataset:stanfordnlp/imdb, arxiv:1909.08593, base_model:EleutherAI/pythia-160m, base_model:finetune:EleutherAI/pythia-160m”
“Downloads: 0 Likes: 5 Tags: safetensors, region:us”
“Downloads: 38 Likes: 0 Tags: safetensors, gpt2, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 5 Tags: region:us”
“Downloads: 0 Likes: 5 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 21822 Likes: 15 Tags: mlx, safetensors, glm_moe_dsa, text-generation, transformers, conversational, arxiv:2505.02390, base_model:zai-org/GLM-5.2, base_model:quantized:zai-org/GLM-5.2, license:mit, 4-bit, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 2 Tags: region:us”
“Downloads: 0 Likes: 1 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: text-to-speech, tts, speech-synthesis, transformer, pytorch, vits, hifigan, en, dataset:LJSpeech, dataset:LibriTTS, license:mit, region:us”
“Downloads: 61 Likes: 3 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 1 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: safetensors, qwen2, region:us”
“Downloads: 0 Likes: 3 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: safetensors, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 1 Tags: region:us”
“Downloads: 74 Likes: 0 Tags: safetensors, whisper, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: safetensors, region:us”
“Downloads: 0 Likes: 1 Tags: region:us”
“Downloads: 0 Likes: 2 Tags: safetensors, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: transformers, safetensors, qwen3_5, image-text-to-text, qwen3_6, token-efficient, efficient-thinking, conversational, base_model:bottlecapai/ThinkingCap-Qwen3.6-27B, base_model:finetune:bottlecapai/ThinkingCap-Qwen3.6-27B, license:apache-2.0, endpoints_compatible”
“Downloads: 0 Likes: 0 Tags: peft, safetensors, base_model:adapter:Qwen/Qwen3-1.7B, lora, sft, transformers, trl, text-generation, conversational, arxiv:1910.09700, base_model:Qwen/Qwen3-1.7B, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: transformers, safetensors, qwen3, text-generation, conversational, arxiv:1910.09700, text-generation-inference, endpoints_compatible, region:us”
“Downloads: 0 Likes: 0 Tags: safetensors, gguf, qwen3, endpoints_compatible, region:us, conversational”
“Downloads: 0 Likes: 0 Tags: task_categories:robotics, language:en, license:other, size_categories:n<1K, format:parquet, modality:tabular, modality:text, modality:video, library:datasets, library:pandas, library:polars, library:mlcroissant”
“Downloads: 13233 Likes: 4 Tags: region:us”
“Downloads: 134 Likes: 0 Tags: region:us”
“Downloads: 153 Likes: 0 Tags: license:mit, region:us”
“Downloads: 1970 Likes: 5 Tags: license:mit, region:us”
“Downloads: 1143 Likes: 0 Tags: region:us”
“Downloads: 664 Likes: 0 Tags: region:us”
“Downloads: 1292 Likes: 2 Tags: license:mit, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 2573 Likes: 1 Tags: region:us”
“Downloads: 980 Likes: 0 Tags: region:us”
“Downloads: 952 Likes: 0 Tags: region:us”
“Downloads: 785 Likes: 0 Tags: region:us”
“Downloads: 10693 Likes: 12 Tags: region:us”
“Downloads: 1209 Likes: 0 Tags: region:us”
“Downloads: 3114 Likes: 0 Tags: region:us”
“Downloads: 2137 Likes: 0 Tags: region:us”
“Downloads: 920 Likes: 0 Tags: region:us”
“Downloads: 41 Likes: 0 Tags: region:us”
“Downloads: 241 Likes: 0 Tags: region:us”
“Downloads: 94 Likes: 1 Tags: task_categories:other, language:en, language:ko, license:cc-by-4.0, size_categories:n<1K, library:mlcroissant, region:us, local-ai, llm-benchmark, ai-server, measured-data, croissant”
“Downloads: 27688 Likes: 21 Tags: region:us”
“Downloads: 707 Likes: 0 Tags: region:us”
“Downloads: 1115 Likes: 0 Tags: region:us”
“Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.”
“Aliases: CVE-2025-3933, GHSA-37mw-44qp-f5jm Transformers is vulnerable to ReDoS attack through its DonutProcessor class”
“Aliases: CVE-2024-3568, GHSA-37q5-v5qm-c9v8 Transformers Deserialization of Untrusted Data vulnerability”
“Aliases: CVE-2025-3262, GHSA-489j-g2vx-39wf Transformers vulnerable to ReDoS attack through its SETTING_RE variable”
“Aliases: CVE-2025-6921, GHSA-4w7r-h757-3r74 Hugging Face Transformers vulnerable to Regular Expression Denial of Service (ReDoS) in the AdamWeightDecay optimizer”
“Aliases: CVE-2025-6638, GHSA-59p9-h35m-wg4g Hugging Face Transformers is vulnerable to ReDoS through its MarianTokenizer”
“Aliases: CVE-2024-12720, GHSA-6rvg-6v2m-4j46 Transformers Regular Expression Denial of Service (ReDoS) vulnerability”
“Aliases: CVE-2025-5197, GHSA-9356-575x-2w9m Hugging Face Transformers Regular Expression Denial of Service (ReDoS) vulnerability”
“Aliases: CVE-2025-1194, GHSA-fpwr-67px-3qhx Transformers Regular Expression Denial of Service (ReDoS) vulnerability”
“Aliases: CVE-2025-3264, GHSA-jjph-296x-mrcr Transformers vulnerable to ReDoS attack through its get_imports() function”
“Aliases: CVE-2025-3777, GHSA-phhr-52qp-3mj4 Transformers's Improper Input Validation vulnerability can be exploited through username injection”
“Aliases: CVE-2025-3263, GHSA-q2wp-rjmx-x6x9 Transformers's ReDoS vulnerability in get_configuration_file can lead to catastrophic backtracking”
“Aliases: CVE-2025-6051, GHSA-rcv9-qm8p-9p6j Hugging Face Transformers library has Regular Expression Denial of Service”
Decision support, not stock advice. This signal is research with cited evidence — not a recommendation to buy, sell, or hold any security.