`CHECK` fail via inputs in `SparseFillEmptyRowsGrad`

ImpactIf SparseFillEmptyRowsGrad is given empty inputs, TensorFlow will crash.pythonimport tensorflow as tftf.rawops.SparseFillEmptyRowsGrad reverseindexmap=, gradvalues=, name=None PatchesWe have patched the issue in GitHub commit af4a6a3c8b95022c351edae94560acc61253a1b8.The fix will be included...

`CHECK` fail in `BCast` overflow

ImpactIf BCast::ToShape is given input larger than an int32, it will crash, despite being supposed to handle up to an int64. An example can be seen in tf.experimental.numpy.outer by passing in large input to the input b.pythonimport tensorflow as tfvalue = tf.constantshape=2, 1024, 1024, 1024,...

pymatgen is vulnerable to Regular Expression Denial of Service (ReDoS)

An exponential ReDoS Regular Expression Denial of Service can be triggered in the pymatgen PyPI package, when an attacker is able to supply arbitrary input to the GaussianInput.fromstring method...

Twisted vulnerable to NameVirtualHost Host header injection

When the host header does not match a configured host, twisted.web.vhost.NameVirtualHost will return a NoResource resource which renders the Host header unescaped into the 404 response allowing HTML and script injection.Example configuration:pythonfrom twisted.web.server import Sitefrom...

TensorFlow vulnerable to `CHECK` fail in `CollectiveGather`

ImpactWhen CollectiveGather receives an scalar input input, it gives a CHECK fails that can be used to trigger a denial of service attack.pythonimport tensorflow as tfarg0=1arg1=1arg2=1arg3=1arg4=3, 3,3arg5='auto'arg6=0arg7=''tf.rawops.CollectiveGatherinput=arg0, groupsize=arg1, groupkey=arg2,...

TensorFlow vulnerable to `CHECK` failures in `AvgPool3DGrad`

ImpactThe implementation of AvgPool3DGradOp does not fully validate the input originputshape. This results in an overflow that results in a CHECK failure which can be used to trigger a denial of service attack:pythonimport tensorflow as tfksize = 1, 1, 1, 1, 1strides = 1, 1, 1, 1, 1padding =...

TensorFlow vulnerable to Int overflow in `RaggedRangeOp`

ImpactThe RaggedRangOp function takes an argument limits that is eventually used to construct a TensorShape as an int64. If limits is a very large float, it can overflow when converted to an int64. This triggers an InvalidArgument but also throws an abort signal that crashes the...

TensorFlow vulnerable to integer overflow in math ops

ImpactWhen RangeSize receives values that do not fit into an int64t, it crashes.cpp auto size = std::isintegral::value ? Eigen::numext::abslimit - start + Eigen::numext::absdelta - T1 / Eigen::numext::absdelta : Eigen::numext::ceil Eigen::numext::abslimit - start / delta; // This check does not...

TensorFlow vulnerable to `CHECK` fail in `FakeQuantWithMinMaxVarsGradient`

ImpactWhen tf.quantization.fakequantwithminmaxvarsgradient receives input min or max that is nonscalar, it gives a CHECK fail that can trigger a denial of service attack.pythonimport tensorflow as tfimport numpy as np arg0=tf.constantvalue=np.random.randomsize=2, 2, shape=2, 2,...

PYSEC-2026-2087

The Config API in Apache Airflow surfaced per-key secrets-backend overrides environment variables like AIRFLOWSECRETSBACKENDKWARGSECRETID and AIRFLOWWORKERSSECRETSBACKENDKWARGSECRETID as synthetic config options whose option names were not in sensitiveconfigvalues, so the masker did not redact...

PYSEC-2026-2303

vLLM is a library for LLM inference and serving. From 0.12.0 to before 0.24.0, sending a pure prompt embeds payload in a /v1/completions request with a model using M-RoPE causes EngineCore to fail an assertion and fatally crash, shutting down the entire server application. Any remote user who is...

openstack-barbican Denial of Service vulnerability

An authorization flaw was found in openstack-barbican, where anyone with an admin role could add secrets to a different project container. This flaw allows an attacker on the network to consume protected resources and cause a denial of service...

Matrix Sydent mishandles emails

util/emailutils.py in Matrix Sydent before 1.0.2 mishandles registration restrictions that are based on e-mail domain, if the allowedlocal3pids option is enabled. This occurs because of potentially unwanted behavior in Python, in which an email.utils.parseaddr call on...

OpenStack Image Registry and Delivery Service (Glance) Improper Input Validation vulnerability

The Sheepdog backend in OpenStack Image Registry and Delivery Service Glance 2013.2 before 2013.2.4 and icehouse before icehouse-rc2 allows remote authenticated users with permission to insert or modify an image to execute arbitrary commands via a crafted location...

Web2py Cross-Site Request Forgery vulnerability

Web2py versions 2.14.5 and below was affected by CSRF Cross Site Request Forgery vulnerability, which allows an attacker to trick a logged-in administrator into performing unwanted actions i.e An attacker can trick a victim into disable the installed application just by visiting a URL...

Matrix Synapse DoS

Matrix Synapse before 0.28.1 is prone to a denial of service flaw where malicious events injected with depth = 263 - 1 render rooms unusable, related to federation/federationbase.py and handlers/message.py, as exploited in the wild in April 2018...

simplejson before 2.6.1 vulnerable to array index error

Array index error in the scanstring function in the json module in Python 2.7 through 3.5 and simplejson before 2.6.1 allows context-dependent attackers to read arbitrary process memory via a negative index value in the idx argument to the rawdecode function...

Mercurial Directory traversal vulnerability

Directory traversal vulnerability in patch.py in Mercurial before 1.0.2 allows user-assisted attackers to modify arbitrary files via ".." dot dot sequences in a patch file...

MoinMoin Multiple cross-site scripting (XSS) vulnerabilities

Multiple cross-site scripting XSS vulnerabilities in MoinMoin 1.5.8 and earlier allow remote attackers to inject arbitrary web script or HTML via 1 certain input processed by formatter/textgedit.py aka the gui editor formatter; 2 a page name, which triggers an injection in PageEditor.py when the...

MoinMoin Cross-Site Scripting (XSS) vulnerability via hitcounts and general parameters

Multiple cross-site scripting XSS vulnerabilities in Info pages in MoinMoin 1.5.7 allow remote attackers to inject arbitrary web script or HTML via the 1 hitcounts and 2 general parameters, different vectors than CVE-2007-0857. NOTE: the provenance of this information is unknown; the details are...

calibre-web is vulnerable to Cross-Site Request Forgery (CSRF)

calibre-web is vulnerable to Cross-Site Request Forgery CSRF...

Malicious code in magique (PyPI)

Part of the "Hades" wave of the Shai-Hulud supply-chain campaign. On 2026-06-08,malicious phantom releases of magique were published to PyPI using stolencredentials. The package executes a bundled JavaScript payload via the Bunruntime on import that harvests and exfiltrates credentials and...

Malicious code in funcdesc (PyPI)

Part of the "Hades" wave of the Shai-Hulud supply-chain campaign. On 2026-06-08,malicious phantom releases of funcdesc were published to PyPI using stolencredentials. The package executes a bundled JavaScript payload via the Bunruntime on import that harvests and exfiltrates credentials and...

Malicious code in executor-http (PyPI)

Part of the "Hades" wave of the Shai-Hulud supply-chain campaign. On 2026-06-08,malicious phantom releases of executor-http were published to PyPI using stolencredentials. The package executes a bundled JavaScript payload via the Bunruntime on import that harvests and exfiltrates credentials and...

Malicious code in coolbox (PyPI)

Part of the "Hades" wave of the Shai-Hulud supply-chain campaign. On 2026-06-08,malicious phantom releases of coolbox were published to PyPI using stolencredentials. The package executes a bundled JavaScript payload via the Bunruntime on import that harvests and exfiltrates credentials and...

Backpropagate: backprop ui --auth and backprop ui --share do not enforce authentication

SummaryIn backpropagate = 1.1.0, the optional Reflex web UI pip install backpropagateui, launched via backprop ui exposes a training control plane: dataset upload, model load, training start/stop, multi-run orchestration, GGUF export, and HuggingFace Hub push.The CLI accepts two operator-facing...

Jupyter Enterprise Gateway: Jinja2 Template Server Side Template Injection resulting in Remote Code Execution

SummaryThe environment variables KERNELXXX used during the rendering of the Kubernetes manifest are vulnerable to Server Side Template Injection SSTI.By including Jinja2 template expressions it is possible to execution Python code and OS Commands in the Enterprise Gateway service.The code can use...

python-statemachine SCXML <data expr> Eval Injection

Summarypython-statemachine 3.1.2 evaluates attributes in SCXML documents using Python's eval. Any application that passes attacker-controlled SCXML content to SCXMLProcessor is vulnerable to arbitrary code execution in the context of the hosting process. DetailsSCXMLProcessor.parsescxmlfile...

BoxLite: Permission Bypass Allows Modification of Read-Only Files

SummaryBoxlite is a sandbox service that allows users to create lightweight virtual machines Boxes and launch OCI containers within them to run untrusted code.One of the core security features claimed by Boxlite is the ability to mount host directories in read-only mode readonly=True into the VM...

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...

Mesop Affected by Unauthenticated Remote Code Execution via Test Suite Route /exec-py

SummaryAn explicit web endpoint inside the ai/ testing module infrastructure directly ingests untrusted Python code strings unconditionally without authentication measures, yielding standard Unrestricted Remote Code Execution. Any individual capable of routing HTTP logic to this server block will...

External Control of File Name or Path in h2oai/h2o-3

Remote unauthenticated attackers can overwrite arbitrary server files with attacker-controllable data. The data that the attacker can control is not entirely arbitrary. h2o writes a CSV/XLS/etc file to disk, so the attacker data is wrapped in quotations and starts with "C1", if they're exporting ...

qdrant input validation failure

qdrant/qdrant version 1.9.0-dev is vulnerable to arbitrary file read and write during the snapshot recovery process. Attackers can exploit this vulnerability by manipulating snapshot files to include symlinks, leading to arbitrary file read by adding a symlink that points to a desired file on the...

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...

llama-index-core Command Injection vulnerability

A command injection vulnerability exists in the run-llama/llamaindex repository, specifically within the safeeval function. Attackers can bypass the intended security mechanism, which checks for the presence of underscores in code generated by LLM, to execute arbitrary code. This is achieved by...

llama-index-core Prompt Injection vulnerability leading to Arbitrary Code Execution

A vulnerability was identified in the executils class of the llamaindex package, specifically within the safeeval function, allowing for prompt injection leading to arbitrary code execution. This issue arises due to insufficient validation of input, which can be exploited to bypass method...

Hugging Face smolagents: Unsafe deserialization in Remote Python Executor leads to RCE

Hugging Face smolagents Remote Python Executor Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face smolagents. Authentication is not required to exploit this...

ExecuTorch integer overflow vulnerability

An integer overflow vulnerability in the loading of ExecuTorch models can cause overlapping allocations, potentially resulting in code execution or other undesirable effects. This issue affects ExecuTorch prior to commit d158236b1dc84539c1b16843bc74054c9dcba006...

BentoML SSRF Vulnerability in File Upload Processing

DescriptionThere's an SSRF in the file upload processing system that allows remote attackers to make arbitrary HTTP requests from the server without authentication. The vulnerability exists in the serialization/deserialization handlers for multipart form data and JSON requests, which automaticall...

AgentScope path traversal vulnerability in save-workflow

A path traversal vulnerability exists in the save-workflow and load-workflow functionality of modelscope/agentscope versions prior to the fix. This vulnerability allows an attacker to read and write arbitrary JSON files on the filesystem, potentially leading to the exposure or modification of...

AgentScope Deserialization Vulnerability

A vulnerability in the RpcAgentServerLauncher class of modelscope/agentscope v0.0.6a3 allows for remote code execution RCE via deserialization of untrusted data using the dill library. The issue occurs in the AgentServerServicer.createagent method, where serialized input is deserialized using...

WMAgent arbitrary code execution via a crafted dbs-client package

WMAgent v1.3.3rc2 and 1.3.3rc1, reqmgr2 1.4.1rc5 and 1.4.0rc2, reqmon 1.4.1rc5, and global-workqueue 1.4.1rc5 allows attackers to execute arbitrary code via a crafted dbs-client package...

OpenStack Object Storage (swift) Code Injection vulnerability

OpenStack Object Storage swift before 1.7.0 uses the loads function in the pickle Python module unsafely when storing and loading metadata in memcached, which allows remote attackers to execute arbitrary code via a crafted pickle object...

PYSEC-2026-244

Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to 1.10.0, the "Shareable Playground" or "Public Flows" in code contains a potential arbitrary file-read vulnerability, depending on the exact flow configuration used. By making a flow public, public execution of...

PYSEC-0000-CVE-2026-56258

Crawl4AI before 0.8.8 contains an arbitrary file write vulnerability in the screenshot and PDF endpoints that allows unauthenticated attackers to write files outside the intended directory via symlink and time-of-check-time-of-use TOCTOU attacks on the outputpath parameter. Remote attackers can...

PYSEC-2026-2111

AIOHTTP is an asynchronous HTTP client/server framework for asyncio and Python. Prior to 3.14.1, during cleanup it is possible for a compressed request body to be decompressed into memory in one chunk. An attacker may be able to send a compressed payload in specific situations that could be...

PYSEC-2026-2191

kafka-python prior to 2.3.2 contains a denial-of-service vulnerability in SCRAM authentication handling that allows a malicious or machine-in-the-middle broker to freeze the client event loop by supplying an excessively large iteration count. In scram.py, ScramClient.processserverfirstmessage...

PYSEC-2026-599

An issue was discovered in OpenStack Keystone before 29.0.2. The Keystone application credential authentication plugin does not verify that the user supplied in the authentication request matches the owner of the application credential. An attacker can authenticate with their own application...

PYSEC-2026-194

Synapse is an open source Matrix homeserver implementation. Prior to 1.152.1, in federated rooms, malicious homeservers can craft room events in such a way that prevents Synapse from providing full history to paginating clients. Clients could therefore fail to display room history. This...

PYSEC-0000-CVE-2026-48522

PyJWT is a JSON Web Token implementation in Python. Prior to 2.13.0, PyJWKClient passes its uri argument directly to urllib.request.urlopen which uses Python stdlib's default OpenerDirector registering HTTPHandler, HTTPSHandler, FTPHandler, FileHandler, and DataHandler. There is currently no...