TensorFlow vulnerable to `CHECK` fail in `TensorListScatter` and `TensorListScatterV2`

ImpactWhen TensorListScatter and TensorListScatterV2 receive an elementshape of a rank greater than one, they give a CHECK fail that can trigger a denial of service attack.pythonimport tensorflow as tfarg0=tf.random.uniformshape=2, 2, 2, dtype=tf.float16, maxval=Nonearg1=tf.random.uniformshape=2,...

TensorFlow vulnerable to `CHECK` fail in `AudioSummaryV2`

ImpactWhen AudioSummaryV2 receives an input samplerate with more than one element, it gives a CHECK fails that can be used to trigger a denial of service attack.pythonimport tensorflow as tfarg0=''arg1=tf.random.uniformshape=1,1, dtype=tf.float32, maxval=Nonearg2=tf.random.uniformshape=2,1,...

TensorFlow vulnerable to `CHECK` fail in `tf.random.gamma`

ImpactWhen tf.random.gamma receives large input shape and rates, it gives a CHECK fail that can trigger a denial of service attack.pythonimport tensorflow as tfarg0=tf.random.uniformshape=4,, dtype=tf.int32, maxval=65536arg1=tf.random.uniformshape=4, 4, dtype=tf.float64,...

TensorFlow vulnerable to `CHECK` fail in `FakeQuantWithMinMaxVarsPerChannel`

ImpactIf FakeQuantWithMinMaxVarsPerChannel is given min or max tensors of a rank other than one, it results in a CHECK fail that can be used to trigger a denial of service attack.pythonimport tensorflow as tfnumbits = 8narrowrange = Falseinputs = tf.constant0, shape=4, dtype=tf.float32min =...

TensorFlow segfault TFLite converter on per-channel quantized transposed convolutions

ImpactWhen converting transposed convolutions using per-channel weight quantization the converter segfaults and crashes the Python process.pythonimport tensorflow as tfclass QuantConv2DTransposedtf.keras.layers.Layer: def buildself, inputshape: self.kernel = self.addweight"kernel", 3, 3,...

PYSEC-2026-2083

Before apache-airflow 3.3.0, a user authorized to read one Dag could disclose the source of other Dags co-located in the same source file. GET /api/v2/dagSources/dagid — and the equivalent Dag-source view in the UI — returned the entire source file without redacting Dags the caller was not...

PYSEC-2026-2140

Crawl4AI is an open-source LLM-friendly web crawler and scraper. Prior to 0.9.0, the Docker API server applied its SSRF destination check on the non-streaming /crawl path but not on the streaming path. handlestreamcrawlrequest passed seed URLs straight to the crawler with no destination validatio...

PYSEC-2026-2139

Crawl4AI is an open-source LLM-friendly web crawler and scraper. Prior to 0.9.0, the Docker API server accepted request-supplied browserconfig.extraargs, which flowed into Chromium's launch arguments. An attacker could inject Chromium switches that replace a child-process launch command together...

PYSEC-2026-2305

vLLM is an inference and serving engine for large language models. From 0.22.0 to 0.23.0, the /v1/audio/transcriptions and /v1/audio/translations routes call request.file.read to fully materialize an uploaded audio file into memory before vLLM checks the documented VLLMMAXAUDIOCLIPFILESIZEMB...

PYSEC-2026-2256

Pillow is a Python imaging library. Prior to 12.3.0, PIL/GdImageFile.py GdImageFile.open read image dimensions from the GD 2.x header and stored them in self.size without calling Image.decompressionbombcheck, allowing a crafted .gd file to trigger excessive C-heap allocation when loaded. This iss...

TensorFlow vulnerable to `CHECK` failure in `SobolSample` via missing validation

ImpactThe implementation of SobolSampleOp is vulnerable to a denial of service via CHECK-failure assertion failure caused by assuming input0, input1, and input2 to be scalar.pythonimport tensorflow as tftf.rawops.SobolSampledim=tf.constant1,0, numresults=tf.constant1, skip=tf.constant1 PatchesWe...

Potential double free of buffer during string decoding

ImpactWhat kind of vulnerability is it? Who is impacted?When an error occurs while reallocating the buffer for string decoding, the buffer gets freed twice.Due to how UltraJSON uses the internal decoder, this double free is impossible to trigger from Python. PatchesHas the problem been patched?...

Missing validation crashes `QuantizeAndDequantizeV4Grad`

ImpactThe implementation of tf.rawops.QuantizeAndDequantizeV4Grad does not fully validate the input arguments. This results in a CHECK-failure which can be used to trigger a denial of service attack:pythonimport tensorflow as tftf.rawops.QuantizeAndDequantizeV4Grad gradients=tf.constant1,...

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 Compute (Nova) allows remote authenticated users to gain privileges via API requests

The Nova EC2 API security group implementation in OpenStack Compute Nova 2013.1 before 2013.2.4 and icehouse before icehouse-rc2 does not enforce RBAC policies for 1 addrules, 2 removerules, 3 destroy, and other unspecified methods in compute/api.py when using non-default policies, which allows...

OpenStack Object Storage (Swift) Sensitive Data Exposure

OpenStack Object Storage Swift before 2.4.0 allows attackers to obtain sensitive information via a PUT tempurl and a DLO object manifest that references an object in another container...

OpenStack Cinder, Glance, and Nova contain Uncontrolled Resource Consumption

The image parser in OpenStack Cinder prior to 7.0.2, and 8.0.0 and above, prior to 9.0.0; Glance prior to 14.00; and Nova prior to 12.0.4 does not properly limit qemu-img calls, which might allow attackers to cause a denial of service memory and disk consumption via a crafted disk image. This iss...

Openstack DBaaS (Trove) Improper Link Resolution Before File Access

The writeconfig function in trove/guestagent/datastore/experimental/mongodb/service.py, resetconfiguration function in trove/guestagent/datastore/experimental/postgresql/service/config.py, writeconfig function in trove/guestagent/datastore/experimental/redis/service.py, writemycnf function in...

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...

keycloak-httpd-client-install symlink attack vulnerability

keycloak-httpd-client-install versions before 0.8 insecurely creates temporary file allowing local attackers to overwrite other files via symbolic link...

keycloak-httpd-client-install Insecure Secrets

keycloak-httpd-client-install versions before 0.8 allow users to insecurely pass password through command line, leaking it via command history and process info to other local users...

PYSEC-2026-2150

In Trail of Bits fickling versions up to and including 0.1.11, the UnsafeImportsML analysis pass unconditionally calls AnalysisContext.shortencodenode on every import node it inspects, regardless of whether the import is flagged as unsafe. This call registers the shortened code representation in...

OpenStack Keystone Denial of Service vulnerability via a large HTTP request

OpenStack Keystone Grizzly before 2013.1, Folsom, and possibly earlier allows remote attackers to cause a denial of service CPU and memory consumption via a large HTTP request, as demonstrated by a long tenantname when requesting a token...

Path Traversal in scout-browser

Scout is a Variant Call Format VCF visualization interface. The Pypi package scout-browser is vulnerable to path traversal due to sendfile call in versions prior to 4.52...

Django Improper Access Control

The LazyUser class in the AuthenticationMiddleware for Django 0.95 does not properly cache the user name across requests, which allows remote authenticated users to gain the privileges of a different user...

Out of bounds read in Tensorflow

ImpactThe TFG dialect of TensorFlow MLIR makes several assumptions about the incoming GraphDef before converting it to the MLIR-based dialect.If an attacker changes the SavedModel format on disk to invalidate these assumptions and the GraphDef is then converted to MLIR-based IR then they can caus...

Open Redirect in CPython that affects users of OpenStack Nova

A vulnerability was found in CPython which is used by openstack-nova's console proxy, noVNC. By crafting a malicious URL, noVNC could be made to redirect to any desired URL...

Cross-site Scripting in Ericsson CodeChecker

In Ericsson CodeChecker prior to 6.18.2, a Stored Cross-site scripting XSS vulnerability in the comments component of the reports viewer allows remote attackers to inject arbitrary web script or HTML via the POST JSON data of the /CodeCheckerService API...

Cross-site scripting vulnerability in TinyMCE plugins

ImpactA cross-site scripting XSS vulnerability was discovered in the URL processing logic of the image and link plugins. The vulnerability allowed arbitrary JavaScript execution when updating an image or link using a specially crafted URL. This issue only impacted users while editing and the...

PYSEC-2026-613

Wagtail is an open source content management system built on Django. In versions prior to 7.0.8, 7.3.3 and 7.4.2, an authenticated admin user can trigger expensive rendition processing with purposefully crafted filter specs resulting in potentially service degradation. The vulnerability is not...

PYSEC-2026-616

Wagtail is an open source content management system built on Django. In versions prior to 7.0.8, 7.3.3 and 7.4.2, reflected cross-site scripting XSS vulnerability exists on the dynamic image URL generator view within the Wagtail admin interface. A user with a limited-permission editor account for...

Malicious code in pantheon-toolsets (PyPI)

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

Malicious code in nucbox (PyPI)

Part of the "Hades" wave of the Shai-Hulud supply-chain campaign. On 2026-06-08,malicious phantom releases of nucbox 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 napari-ufish (PyPI)

Part of the "Hades" wave of the Shai-Hulud supply-chain campaign. On 2026-06-08,malicious phantom releases of napari-ufish 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 mrbios (PyPI)

Part of the "Hades" wave of the Shai-Hulud supply-chain campaign. On 2026-06-08,malicious phantom releases of mrbios 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 nhmpy (PyPI)

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

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...

praisonai-platform: Any workspace member can add arbitrary user as owner via POST /workspaces/{id}/members

SummaryType: Privilege escalation / cross-tenant member injection. The POST /workspaces/workspaceid/members endpoint is gated only by requireworkspacememberworkspaceid default minrole="member" and forwards the request body's userid and role straight into MemberService.addworkspaceid, userid, role...

SGLang's multimodal generation runtime has an unauthenticated path traversal vulnerability

SGLang's multimodal generation runtime is vulnerable to an unauthenticated path traversal vulnerability, allowing an attacker to write arbitrary files anywhere the server process has write access, by including ../ sequences in the upload filename when sent to specific endpoints...

Horovod contains an insecure deserialization vulnerability in its KVStore HTTP server component

Horovod thru 0.28.1 contains an insecure deserialization vulnerability CWE-502 in its KVStore HTTP server component. The KVStore server, used for distributed task coordination, lacks authentication and authorization controls, allowing any remote attacker to write arbitrary data via HTTP PUT...

SGLang: Unauthenticated RCE via --enable-custom-logit-processor

SGLang's multimodal generation runtime is vulnerable to unauthenticated remote code execution when the --enable-custom-logit-processor option is enabled, as Python objects loaded via dill.loads will be deserialized without validation...

llm CLI tool contains a code injection vulnerability via `--functions` command-line argument

The llm CLI tool thru 0.27.1 contains a critical code injection vulnerability via its --functions command-line argument. This argument is intended to allow users to provide custom Python function definitions. However, the tool directly executes the provided code using the unsafe exec function...

PraisonAI call server exposes unauthenticated agent listing, invocation, and deletion when CALL_SERVER_TOKEN is unset

SummaryPraisonAI's call server exposes a network-facing agent control API without authentication when CALLSERVERTOKEN is not configured.The affected component is the praisonai.api.agentinvoke router as mounted by praisonai.api.call. The authentication helper verifytoken fails open when...

amazon-redshift-python-driver vulnerable to Remote Code Execution via eval() Injection

Summaryamazon-redshift-python-driver is the official Python connector for Amazon Redshift. In versions 2.1.13 and earlier, the driver insufficiently validates data received from the server during query result processing. A rogue server or man-in-the-middle could leverage this to execute arbitrary...

wger: cross-tenant password reset and plaintext disclosure via gym=None bypass

SummaryThe resetuserpassword and gympermissionsuseredit views in wger perform a gym-scope authorization check using Python object comparison != that evaluates None != None as False, silently bypassing the guard when both the attacker and victim have no gym assignment gym=None. A user with...

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...

Marimo: Pre-Auth Remote Code Execution via Terminal WebSocket Authentication Bypass

SummaryMarimo 19.6k stars has a Pre-Auth RCE vulnerability. The terminal WebSocket endpoint /terminal/ws lacks authentication validation, allowing an unauthenticated attacker to obtain a full PTY shell and execute arbitrary system commands.Unlike other WebSocket endpoints e.g., /ws that correctly...

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 ...