PYSEC-2026-406 mamba language model framework vulnerable to insecure deserialization when loading pre-trained models from HuggingFace Hub

The mamba language model framework thru 2.2.6 is vulnerable to insecure deserialization CWE-502 when loading pre-trained models from HuggingFace Hub. The MambaLMHeadModel.frompretrained method uses torch.load to load the pytorchmodel.bin weight file without enabling the security-restrictive...

PYSEC-2026-405 Ludwig framework is vulnerable to insecure deserialization in its model serving component

The Ludwig framework thru 0.10.4 is vulnerable to insecure deserialization CWE-502 in its model serving component. When starting a model server with the ludwig serve command, the framework loads model weight files using torch.load without enabling the security-restrictive weightsonly=True...

PraisonAI's unauthenticated A2A official example can reach real LLM-driven `eval()` tool execution

SummaryThe first-party PraisonAI A2A server example combines three behaviors into a remotely exploitable Critical chain:1. The example exposes an A2A server without configuring authtoken.2. The same example binds the server to 0.0.0.0.3. The example registers a calculateexpression tool implemente...

Ludwig framework is vulnerable to insecure deserialization in its model serving component

The Ludwig framework thru 0.10.4 is vulnerable to insecure deserialization CWE-502 in its model serving component. When starting a model server with the ludwig serve command, the framework loads model weight files using torch.load without enabling the security-restrictive weightsonly=True...

PYSEC-2026-478 PraisonAI's unauthenticated A2A official example can reach real LLM-driven `eval()` tool execution

Summary The first-party PraisonAI A2A server example combines three behaviors into a remotely exploitable Critical chain: 1. The example exposes an A2A server without configuring authtoken. 2. The same example binds the server to 0.0.0.0. 3. The example registers a calculateexpression tool...

PYSEC-2026-461 PraisonAI Vulnerable to OS Command Injection

The executecommand function and workflow shell execution are exposed to user-controlled input via agent workflows, YAML definitions, and LLM-generated tool calls, allowing attackers to inject arbitrary shell commands through shell metacharacters. --- Description PraisonAI's workflow system and...

PYSEC-2026-256 Agno is vulnerable to Eval Injection

Agno versions prior to 2.3.24 contain an arbitrary code execution vulnerability in the model execution component that allows attackers to execute arbitrary Python code by manipulating the fieldtype parameter passed to eval. Attackers can influence the fieldtype value in a FunctionCall to achieve...

PYSEC-2026-423 MLflow Command Injection vulnerability

A command injection vulnerability exists in MLflow's model serving container initialization code, specifically in the installmodeldependenciestoenv function. When deploying a model with envmanager=LOCAL, MLflow reads dependency specifications from the model artifact's pythonenv.yaml file and...

MLflow Command Injection vulnerability

A command injection vulnerability exists in MLflow's model serving container initialization code, specifically in the installmodeldependenciestoenv function. When deploying a model with envmanager=LOCAL, MLflow reads dependency specifications from the model artifact's pythonenv.yaml file and...

Mlflow: Command Injection when serving models with enable_mlserver=True

A command injection vulnerability exists in Mlflow when serving a model with enablemlserver=True. The modeluri is embedded directly into a shell command executed via bash -c without proper sanitization. If the modeluri contains shell metacharacters, such as $ or backticks, it allows for command...

PYSEC-2026-424 Mlflow: Command Injection when serving models with enable_mlserver=True

A command injection vulnerability exists in Mlflow when serving a model with enablemlserver=True. The modeluri is embedded directly into a shell command executed via bash -c without proper sanitization. If the modeluri contains shell metacharacters, such as $ or backticks, it allows for command...

PYSEC-2026-553 TorchServe Server-Side Request Forgery vulnerability

Impact Remote Server-Side Request Forgery SSRF Issue: TorchServe default configuration lacks proper input validation, enabling third parties to invoke remote HTTP download requests and write files to the disk. This issue could be taken advantage of to compromise the integrity of the system and...

vanna vulnerable to remote code execution caused by prompt injection

In the latest version of vanna-ai/vanna, the vanna.ask function is vulnerable to remote code execution due to prompt injection. The root cause is the lack of a sandbox when executing LLM-generated code, allowing an attacker to manipulate the code executed by the exec function in...

litellm vulnerable to remote code execution based on using eval unsafely

BerriAI/litellm version v1.35.8 contains a vulnerability where an attacker can achieve remote code execution. The vulnerability exists in the adddeployment function, which decodes and decrypts environment variables from base64 and assigns them to os.environ. An attacker can exploit this by sendin...

TorchServe vulnerable to bypass of allowed_urls configuration

ImpactTorchServe's check on allowedurls configuration can be by-passed if the URL contains characters such as ".." but it does not prevent the model from being downloaded into the model store. Once a file is downloaded, it can be referenced without providing a URL the second time, which effective...

PYSEC-2026-392 llama-cpp-python vulnerable to Remote Code Execution by Server-Side Template Injection in Model Metadata

Description llama-cpp-python depends on class Llama in llama.py to load .gguf llama.cpp or Latency Machine Learning Models. The init constructor built in the Llama takes several parameters to configure the loading and running of the model. Other than NUMA, LoRa settings, loading tokenizers, and...

llama-cpp-python vulnerable to Remote Code Execution by Server-Side Template Injection in Model Metadata

Descriptionllama-cpp-python depends on class Llama in llama.py to load .gguf llama.cpp or Latency Machine Learning Models. The init constructor built in the Llama takes several parameters to configure the loading and running of the model. Other than NUMA, LoRa settings, loading tokenizers, and...

Keras code injection vulnerability

A arbitrary code injection vulnerability in TensorFlow's Keras framework 2.13 allows attackers to execute arbitrary code with the same permissions as the application using a model that allow arbitrary code irrespective of the application...

Rasa Allows Remote Code Execution via Remote Model Loading

VulnerabilityA vulnerability has been identified in Rasa Pro and Rasa Open Source that enables an attacker who has the ability to load a maliciously crafted model remotely into a Rasa instance to achieve Remote Code Execution.The prerequisites for this are:- The HTTP API must be enabled on the Ra...

Rasa Allows Remote Code Execution via Remote Model Loading

VulnerabilityA vulnerability has been identified in Rasa Pro and Rasa Open Source that enables an attacker who has the ability to load a maliciously crafted model remotely into a Rasa instance to achieve Remote Code Execution.The prerequisites for this are:- The HTTP API must be enabled on the Ra...