Browse / Unsafe Model Loading and Supply Chain
Unsafe Model Loading and Supply Chain
Malicious model files, pickle deserialisation, poisoned packages and artifacts. 241 records. Reference: OWASP LLM03 · ATLAS AML.T0010.
- Unsafe deserialisation 92
- Code and command execution 59
- Denial of service 23
- Memory corruption 19
- Other weaknesses 17
- Path traversal and file access 8
- Publish Poisoned AI Artifacts: Models 5
- Cross-site scripting 4
- Authentication and access control 4
- Data exfiltration 3
- Supply Chain Compromise: Software 3
- Supply Chain Compromise: Data Repository 2
- Information exposure 1
- Publish Poisoned AI Artifacts: AI Agent Tools 1
- ONNX: Out-of-bounds read in ConvTranspose shape inference via undersized weight rank A crafted model whose weight tensor has fewer than three dimensions triggers an out-of-bounds read during shape inference, crashing the process.
- ONNX: Null Pointer Dereference in ScanInferenceFunction Due to Missing Required Attribute
- ONNX: Denial of service via unchecked expansion of packed sub-byte tensors in ONNX Python APIs
- vLLM: Public multimodal SHM handle forgery reaches worker-side `pickle.loads()` with insecure serialization disabled Forged shared-memory handles in vLLM's multimodal cache reach pickle.loads in the worker, and this works without the insecure-serialisation opt-in.
- NLTK: WekaClassifier passes an unvalidated caller-controlled model path to the weka JVM, bypassing nltk.pathsec containment
- PraisonAI: Unsafe Dynamic Module Loading Leads to Arbitrary Code Execution via tools.py in AgentFlow A workflow directory containing a malicious tools.py is imported and executed without sandboxing when a step resolves a structured output type, ignoring the tool allow setting.
- Docling imports plugin entry points before the allow_external_plugins check Docling imports every installed plugin module before applying the external-plugin restriction, so third-party import-time code runs even when external plugins are disabled.
- NLTK: Missing Post-Download Integrity Verification Allows Malicious Package Injection The NLTK downloader unpacks downloaded archives without checking their hash, so a tampered package can deliver malicious pickle files.
- NLTK: Unsafe Pickle Deserialization in TransitionParser Allows Remote Code Execution
- NLTK: Allowlisted pickle loaders still permit code execution in current source
- vLLM: Incomplete CVE-2025-62164 remediation can be bypassed by concurrent prompt parts
- NLTK: Default ENFORCE=False Disables All pathsec Security Controls
- MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False safety control bypassed by mlflow.statsmodels flavor — RCE via crafted model artifact
- NLTK AllowlistUnpickler dotted-name validation bypass allows remote code execution
- Accelerate path traversal and denial of service via sharded checkpoint weight_map entries A sharded checkpoint index with traversal or absolute paths, or named pipes, makes the loader read arbitrary files or hang while loading weights.
- Feast vulnerable to arbitrary code execution via unsafe deserialization of dill-serialized UDFs
- Keras model loading is vulnerable to denial of service through HDF5 shape bombs A specially crafted .keras archive declares oversized HDF5 datasets, so loading it allocates unbounded memory and the process is killed.
- Flowise: Remote Code Execution Vulnerability in CSVAgent
- ONNX: Heap-Buffer-Overflow READ in Gemm Version Converter Adapter via Undersized Input Shape
- Ray: Arbitrary code execution via ray.data.read_webdataset default decoder: pickle.loads(value) and torch.load(weights_only=False)
- Keras: TorchModuleWrapper can deserialize unsafe PyTorch pickle data
- PyTorch Lightning allows arbitrary code execution through checkpoint _instantiator hyperparameters
- ONNX has Null Pointer Dereference in Upsample Version Converter Adapter (Zero Inputs)
- NLTK Stanford wrapper classes execute untrusted JAR files without verification Five Stanford wrapper classes run user-supplied JAR files through Java without the SHA256 verification added to one sibling class.
- Keras: Lambda deserialization can bypass safe mode and execute code
- Keras: HDF5 virtual datasets can disclose local files A malicious .keras or .h5 model file with virtual datasets pointing at local HDF5 files makes Keras read those files when the model is loaded, because the earlier fix missed this check.
- LangGraph Checkpoint: Unsafe JSON deserialization in checkpoint loading Tampered checkpoint bytes in a persistent store can make the JSON serializer rebuild arbitrary Python objects when a checkpoint is loaded.
- Feast vulnerable to unsafe deserialization RCE in registry server via UDF proto
- Open WebUI: Forged model meta.knowledge allows cross-user file read and deletion
- vLLM: Security Check Bypass via assert Statement in Activation Function Loading Allows Arbitrary Code Execution
- ChromaDB has a code injection vulnerability A model repository with remote code enabled is pulled into a collection update call, letting a user with update permission execute code on the server.
- vLLM's Artifact Pin Decay allows pinned deployments to load unpinned code, weights, and processors Revision pins apply to only some artifacts, so dynamic code, processors and side weights can be loaded from an unpinned revision.
- PraisonAI: Arbitrary code execution via unguarded `spec.loader.exec_module` in `agents_generator.py` - sibling of CVE-2026-44334 Module paths read from YAML configuration are executed through importlib at call sites missed by the earlier environment gate, running attacker-supplied Python code.
- HuggingFace transformers vulnerable to remote code execution A crafted config.json points an internal attention field at a hub repo, so from_pretrained fetches and runs attacker Python without trust_remote_code.
- Cleartext storage of HMAC signing key in Amazon SageMaker Python SDK ModelBuilder/Serve path A signing key readable through cloud describe APIs lets an attacker forge integrity signatures so malicious model artifacts are accepted and run in the serving container.
- Amazon SageMaker Python SDK is missing integrity verification in its Triton inference handler
- Diffusers: TOCTOU Trust Remote Code Bypass A race between the trust check and the load step lets a Hub repository's custom pipeline code run without the trust_remote_code flag.
- ChromaDB Python project has a pre-authentication code injection vulnerability Sending a model repository with trust_remote_code enabled to a collection endpoint makes the server run attacker-supplied Python without authentication.
- MLFlow Creates a Temporary File With Insecure Permissions Model artifact temp directories are created world- or group-writable, so local users can replace cloudpickle files that are later loaded and executed.
- SGLang: Unauthenticated RCE via --enable-custom-logit-processor A custom logit processor sent to the server is deserialised with dill.loads, so an unauthenticated request can run Python code when the option is enabled.
- SGLanG: Multimodal scheduler deserializes untrusted pickle data on 0.0.0.0 ROUTER socket
- LangSmith SDK: Public prompt pull deserializes untrusted manifests without trust boundary warning Pulling a public prompt by owner/name deserialises its manifest as trusted configuration, so attacker-authored constructor arguments and environment-backed secret references can run or leak.
- Ludwig framework is vulnerable to insecure deserialization through its predict() method.
- Ludwig framework is vulnerable to insecure deserialization in its model serving component
- PyTorch Lightning load_from_checkpoint has an insecure checkpoint deserialization load_from_checkpoint calls torch.load without weights_only, so a malicious checkpoint runs pickle-based code when it is loaded.
- LangChain vulnerable to unsafe deserialization of attacker-controlled objects through overly broad `load()` allowlists
- open-webui Vulnerable to Stored XSS via Model Description
- Diffusers has a `trust_remote_code` bypass via `custom_pipeline` and local custom components Remote code in a custom pipeline or local model folder runs even when trust_remote_code is false, because the check sits in the download step that those paths skip.
- Keras vulnerable to DoS via Malicious .keras Model (HDF5 Shape Bomb Causes Petabyte Allocation in KerasFileEditor)
- SGLang has an Improper Input Validation/Injection Issue A remotely supplied model or tokenizer reference causes unsafe deserialisation inside the transformers helper that loads the tokenizer.