On July 7, 2026, the Open Source Vulnerabilities (OSV) database published a batch of CVEs targeting Hugging Face's...
On July 7, 2026, the Open Source Vulnerabilities (OSV) database published a batch of CVEs targeting Hugging Face's Transformers library.
confidence score
Strong evidence: 3 independent source classes support this read.
signal brief
On July 7, 2026, the Open Source Vulnerabilities (OSV) database published a batch of CVEs targeting Hugging Face's Transformers library. The advisories (PYSEC-2026-1977 through 1988) detail multiple Regular Expression Denial of Service (ReDoS) vulnerabilities and a deserialization flaw. Specific vulnerable components include the DonutProcessor class, SETTING_RE variable, AdamWeightDecay optimizer, MarianTokenizer, get_imports() function, and get_configuration_file. One advisory notes a username injection vulnerability via improper input validation. These vulnerabilities could allow attackers to cause denial of service or execute untrusted code. The disclosures come shortly before the release of transformers 5.14.1 on July 16, suggesting patches may be included. However, the publication of multiple CVEs in a short period erodes trust in the library's security posture and may prompt users to evaluate alternatives or apply urgent updates, impacting HuggingFace's reputation in the AI infrastructure ecosystem.
What the sources said:
- PYSEC-2026-1977: "Transformers is vulnerable to ReDoS attack through its DonutProcessor class"
- PYSEC-2026-1981: "Hugging Face Transformers is vulnerable to ReDoS through its MarianTokenizer"
- PYSEC-2026-1986: "Transformers's Improper Input Validation vulnerability can be exploited through username injection"
source data used
“Downloads: 0 Likes: 0 Tags: transformers, safetensors, generated_from_trainer, sft, trl, unsloth, base_model:huihui-ai/Huihui-Qwen3.5-4B-abliterated, base_model:finetune:huihui-ai/Huihui-Qwen3.5-4B-abliterated, endpoints_compatible, region:us”
“Downloads: 0 Likes: 0 Tags: not-for-all-audiences, region:us”
“Downloads: 0 Likes: 0 Tags: safetensors, detr, region:us”
“Downloads: 0 Likes: 0 Tags: transformers, safetensors, gpt2, text-generation, generated_from_trainer, sft, trl, base_model:goldfish-models/nld_latn_10mb, base_model:finetune:goldfish-models/nld_latn_10mb, text-generation-inference, endpoints_compatible, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 5 Tags: gguf, unsloth, fine tune, heretic, uncensored, abliterated, ara, MTP GGUF Quants, Regular GGUF Quants, qwen3.6, multi-stage tuned, thinking”
“Downloads: 0 Likes: 0 Tags: peft, safetensors, lora, transformers, qwen3, research, control-adapter, en, dataset:brikdavies/dualmsm-finetune-mixtures, base_model:Qwen/Qwen3-14B-Base, base_model:adapter:Qwen/Qwen3-14B-Base, region:us”
“Downloads: 0 Likes: 0 Tags: dots_tts, license:apache-2.0, region:us”
“Downloads: 3626 Likes: 0 Tags: safetensors, qwen3_5, region:us”
“Downloads: 0 Likes: 5 Tags: region:us”
“Downloads: 0 Likes: 1 Tags: region:us”
“Downloads: 535 Likes: 0 Tags: gguf, license:mit, endpoints_compatible, region:us, conversational”
“Downloads: 0 Likes: 0 Tags: dots_tts, license:apache-2.0, region:us”
“Downloads: 0 Likes: 0 Tags: not-for-all-audiences, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: nemo, tensorboard, region:us”
“Downloads: 0 Likes: 0 Tags: diffusers, text-to-image, stable-diffusion, stable-diffusion-xl, base_model:RunDiffusion/Juggernaut-X-Hyper, base_model:finetune:RunDiffusion/Juggernaut-X-Hyper, license:creativeml-openrail-m, endpoints_compatible, diffusers:StableDiffusionXLPipeline, region:us”
“Downloads: 0 Likes: 0 Tags: dots_tts, license:apache-2.0, region:us”
“Downloads: 0 Likes: 3 Tags: region:us”
“Downloads: 0 Likes: 1 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: safetensors, region:us”
“Downloads: 0 Likes: 0 Tags: transformers, safetensors, generated_from_trainer, sft, unsloth, trl, base_model:unsloth/Llama-3.2-3B-Instruct-bnb-4bit, base_model:finetune:unsloth/Llama-3.2-3B-Instruct-bnb-4bit, endpoints_compatible, region:us”
“Downloads: 0 Likes: 5 Tags: transformers, tensorboard, safetensors, ida-family, ida-lattice, causal-lm, ai-body, paired-body, governed-memory, recurrent-state, local-attention, cognitive-routing”
“Downloads: 0 Likes: 0 Tags: diffusers, text-to-image, lora, krea2, template:sd-lora, base_model:krea/Krea-2-Raw, base_model:adapter:krea/Krea-2-Raw, license:apache-2.0, region:us”
“Downloads: 544 Likes: 0 Tags: transformers, tensorboard, safetensors, vision-encoder-decoder, image-text-to-text, generated_from_trainer, base_model:yuchiahung/donut_f21_v4_ep3_edit_distance_loss, base_model:finetune:yuchiahung/donut_f21_v4_ep3_edit_distance_loss, license:mit, endpoints_compatible,”
“Downloads: 0 Likes: 2 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: onnx, region:us”
“Downloads: 0 Likes: 0 Tags: transformers, pytorch, bert, feature-extraction, license:mit, endpoints_compatible, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: openpi, robotics, vla, pi0.5, libero, imitation-learning, license:apache-2.0, region:us”
“Downloads: 0 Likes: 0 Tags: transformers, mathematics, license:apache-2.0, endpoints_compatible, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 1 Tags: peft, onnx, safetensors, lora, onnxruntime-genai, agent-memory, qwen2, base_model:Qwen/Qwen2.5-0.5B-Instruct, base_model:adapter:Qwen/Qwen2.5-0.5B-Instruct, license:apache-2.0, region:us”
“Downloads: 0 Likes: 0 Tags: peft, safetensors, base_model:adapter:lerobot/pi05_base, lora, arxiv:1910.09700, region:us”
“Downloads: 0 Likes: 0 Tags: yoga, posture-correction, pose-estimation, pytorch, resmlp, gru, attention, computer-vision, mediapipe, tabular-classification, en, dataset:custom”
“Downloads: 0 Likes: 0 Tags: gguf, llama.cpp, minimax, minimax-m3, moe, msa, q4_k_m, roleplay, creative-writing, long-context, uncensored, abliterated”
“Downloads: 0 Likes: 0 Tags: CartPole-v1, reinforce, reinforcement-learning, custom-implementation, deep-rl-class, model-index, region:us”
“Downloads: 1301 Likes: 0 Tags: safetensors, qwen3_5, ocr, tibetan, qwen3.5, image-text-to-text, conversational, bo, base_model:Qwen/Qwen3.5-0.8B, base_model:finetune:Qwen/Qwen3.5-0.8B, region:us”
“Downloads: 0 Likes: 0 Tags: transformers, safetensors, qwen2, text-generation, generated_from_trainer, sft, unsloth, trl, conversational, text-generation-inference, endpoints_compatible, region:us”
“Downloads: 0 Likes: 0 Tags: gguf, quantized, llama.cpp, scorecard, governance, validated, local-llm, on-device, agentic, tool-calling, function-calling, agents”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: safetensors, region:us”
“Downloads: 0 Likes: 0 Tags: gguf, qwen3.5, moe, mixture-of-experts, quantized, imatrix, heretic, decensored, abliterated, uncensored, mpoa, mtp”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 2 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: onnx, vits, region:us”
“Downloads: 0 Likes: 0 Tags: onnx, region:us”
“Downloads: 0 Likes: 0 Tags: onnx, vits, region:us”
“Downloads: 0 Likes: 0 Tags: onnx, vits, region:us”
“Downloads: 0 Likes: 0 Tags: safetensors, belief-injection, persuasion, research-only, base_model:Qwen/Qwen3-0.6B, base_model:finetune:Qwen/Qwen3-0.6B, license:apache-2.0, region:us”
“Downloads: 2802 Likes: 2 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: region:us”
“Downloads: 5453 Likes: 5 Tags: license:apache-2.0, region:us”
“Downloads: 0 Likes: 1 Tags: task_categories:text-generation, language:en, language:de, license:cc-by-4.0, size_categories:100K<n<1M, region:us, voice-acting, text-to-speech, dramabox, synthetic, emotion”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 3061 Likes: 0 Tags: license:other, region:us, leaderboard, benchmark, mirror”
“Downloads: 922 Likes: 2 Tags: task_categories:text-generation, task_categories:question-answering, language:en, license:apache-2.0, size_categories:n<1K, region:us, science, research, formal-verification, lean4, benchmark, p2p”
“Downloads: 27771 Likes: 23 Tags: region:us”
“Downloads: 807 Likes: 1 Tags: region:us”
“Downloads: 18 Likes: 0 Tags: license:apache-2.0, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 22772 Likes: 0 Tags: license:unknown, region:us”
“Downloads: 0 Likes: 0 Tags: region:us”
“Downloads: 0 Likes: 0 Tags: license:cc-by-nc-4.0, size_categories:n<1K, modality:image, modality:video, library:datasets, library:mlcroissant, region:us, video, portfolio, scrub, source-material”
“Downloads: 644 Likes: 0 Tags: region:us”
“Downloads: 11082 Likes: 13 Tags: region:us”
“Downloads: 1773 Likes: 1 Tags: license:apache-2.0, region:us”
“Downloads: 5374 Likes: 3 Tags: license:mit, region:us”
“Downloads: 8511 Likes: 2 Tags: region:us”
“Downloads: 6372 Likes: 12 Tags: license:mit, region:us”
“Downloads: 4663 Likes: 1 Tags: region:us”
“Downloads: 2169 Likes: 2 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.