PT-2026-96010

Name of the Vulnerable Software and Affected Versions piskvorky gensim versions prior to 4.4.1 Description A deserialization issue exists in the Model Loader component within the Load function of the gensim/utils.py file. A remote attacker can trigger this by manipulating the fname argument, whic...

CVE-2026-54559

PocketSphinx is a small speech recognizer. Prior to 5.1.1, the trie language-model loaders in src/lm/ngrammodeltrie.c do not adequately validate boundary conditions in ARPA, DMP, and binary format headers, and the acoustic-model loaders in src/mdef.c and src/util/bio.c use sscanf with unbounded...

CVE-2026-90553 vLLM before 0.28.0 Remote Code Execution via LlavaOnevision2 processor

vLLM before 0.28.0 contains a remote code execution vulnerability in the LlavaOnevision2 processor loader that ignores the trustremotecode parameter when loading remote processor classes. Attackers can craft a malicious model with arbitrary code in processingllavaonevision2.py that executes with...

CVE-2026-76841

Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trustremotecode=True as a literal or as an unconditional default: RerankModel.gettokenizer in...

GO-2026-5748 Ollama contains a heap out-of-bounds read vulnerability in the GGUF model loader in github.com/ollama/ollama

Ollama is vulnerable to a heap out-of-bounds read in its GGUF model loader. A specially crafted GGUF file can cause the loader to read beyond the allocated heap buffer, potentially leading to a denial of service crash...

Stanza: Remote Code Execution via Unsafe Pickle Deserialization in Model Loaders

SummaryStanza 1.12.0 attempts to safely load PyTorch checkpoint files using torch.load..., weightsonly=True, but automatically falls back to the fully unsafe torch.load..., weightsonly=False when the safe load raises pickle.UnpicklingError. Because the UnpicklingError condition is fully...

CVE-2026-54499 Stanza: Remote Code Execution via Unsafe Pickle Deserialization in Model Loaders

Stanza is a Stanford NLP Python library for tokenization, sentence segmentation, NER, and parsing of many human languages. Prior to 1.12.2, Stanza model loaders such as stanza.models.common.pretrain.Pretrain.load attempt torch.load..., weightsonly=True but fall back to torch.load...,...

CVE-2026-14604

A vulnerability was determined in Open Asset Import Library Assimp up to 6.0.4. Affected is the function Assimp::Exporter::ExportToBlob of the file code/AssetLib/Ply/PlyLoader.cpp of the component PLY Model Handler. This manipulation causes double free. The attack can be initiated remotely. The...

CVE-2026-10230

A flaw was found in Assimp, specifically within the Half-Life 1 MDL Loader component. A local attacker could exploit a heap-based buffer overflow vulnerability in the readanimations function of HL1MDLLoader.cpp. This could lead to information disclosure, denial of service, or potentially arbitrar...

Out-of-bounds Read

Overview Affected versions of this package are vulnerable to Out-of-bounds Read in the readsequenceinfos function of the Half-Life 1 MDL Loader component when processing the aiString argument. An attacker can access sensitive information by supplying specially crafted input that causes the proces...

Heap-based Buffer Overflow

Overview Affected versions of this package are vulnerable to Heap-based Buffer Overflow in the readmeshes function of the Half-Life 1 MDL Loader component. An attacker can execute arbitrary code or cause a denial of service by providing a specially crafted MDL file that triggers a heap-based buff...

CVE-2026-10233

A security vulnerability has been detected in Assimp up to 6.0.4. Affected by this issue is the function HL1MDLLoader::readsequenceinfos of the file HL1MDLLoader.cpp of the component Half-Life 1 MDL Loader. The manipulation of the argument aiString leads to out-of-bounds read. The attack needs to...

EUVD-2026-33566

A security vulnerability has been detected in Assimp up to 6.0.4. Affected by this issue is the function HL1MDLLoader::readsequenceinfos of the file HL1MDLLoader.cpp of the component Half-Life 1 MDL Loader. The manipulation of the argument aiString leads to out-of-bounds read. The attack needs to...

EUVD-2026-33563

A vulnerability was identified in Assimp up to 6.0.4. This impacts the function Assimp::MDL::HalfLife::HL1MDLLoader::readanimations of the file HL1MDLLoader.cpp of the component Half-Life 1 MDL Loader. Such manipulation leads to heap-based buffer overflow. The attack must be carried out locally...

CVE-2026-10230 Assimp Half-Life 1 MDL Loader HL1MDLLoader.cpp read_animations heap-based overflow

A vulnerability was identified in Assimp up to 6.0.4. This impacts the function Assimp::MDL::HalfLife::HL1MDLLoader::readanimations of the file HL1MDLLoader.cpp of the component Half-Life 1 MDL Loader. Such manipulation leads to heap-based buffer overflow. The attack must be carried out locally...

CVE-2026-10229

A vulnerability was determined in Assimp up to 6.0.4. This affects the function HL1MDLLoader::readmeshes of the file HL1MDLLoader.cpp of the component Half-Life 1 MDL Loader. This manipulation causes heap-based buffer overflow. The attack is restricted to local execution. The exploit has been...

PT-2026-45275

A security flaw has been discovered in Assimp up to 6.0.4. Affected is the function HL1MDLLoader::extract anim value of the file HL1MDLLoader.cpp of the component Half-Life 1 MDL Loader. Performing a manipulation of the argument num.total results in heap-based buffer overflow. The attack must be...

CVE-2026-7482

A flaw was found in Ollama. A remote attacker can exploit a heap out-of-bounds read vulnerability in the GGUF model loader by providing a specially crafted GGUF GGML Unified Format file to the /api/create endpoint. This allows the attacker to read beyond the allocated memory buffer, potentially...

Keras vulnerable to DoS via Malicious .keras Model (HDF5 Shape Bomb Causes Petabyte Allocation in KerasFileEditor)

Summary Keras’s model loader KerasFileEditor unsafely loads user-supplied .keras model files containing HDF5-based weight files without performing any validation on HDF5 dataset metadata. An attacker can craft a .keras archive containing a valid model.weights.h5 file whose dataset declares an...

SUSE CVE-2026-7482

Ollama before 0.17.1 contains a heap out-of-bounds read vulnerability in the GGUF model loader. The /api/create endpoint accepts an attacker-supplied GGUF file in which the declared tensor offset and size exceed the file's actual length; during quantization in fs/ggml/gguf.go and...