Open WebUI: Unapproved accounts can open terminal sessions via a WebSocket auth path missing the role check

SummaryThe terminal WebSocket route authenticates its own first-message JWT instead of going through the HTTP dependency chain, and never applies the role check that getverifieduser enforces on every HTTP terminal route. An account whose role is pending, meaning registered but not approved, or...

Open WebUI: Client-side SSRF via unrestricted external resource loading in Vega/Vega-Lite chart rendering

SummaryOpen WebUI renders vega and vega-lite fenced code blocks in chat content by building a Vega view in the viewer's browser without a restricted resource loader. Any user who can place such a block where another user will see it can make that user's browser issue attacker-chosen outbound GET...

PYSEC-2026-3623

Apache Polaris did not consistently validate storage locations supplied during table and view registration.An authenticated principal with permission to register a table or view could, depending on the affected release and registration path, cause Polaris to use the catalog's storage credentials ...

Penelope unsafe tar extraction allows arbitrary local file write via crafted session archive

SummaryPenelope versions prior to 0.19.3 extracted tar archives received from remote sessions without validating archive member paths. When using the affected Unix download path, a malicious or compromised remote session could return a crafted tar archive containing path traversal entries, such a...

zilliztech deep-searcher has an Incorrect Privilege Assignment issue

A weakness has been identified in zilliztech deep-searcher up to 0.0.2. This affects the function CollectionRouter.invoke of the file deepsearcher/agent/collectionrouter.py. This manipulation of the argument kwargs causes improper access controls. Remote exploitation of the attack is possible. Th...

pypdf: Possible infinite loop for not terminated inline images

ImpactAn attacker who uses this vulnerability can craft a PDF which leads to an infinite loop. This requires parsing the content stream of a page with a not terminated inline image, as done when extracting the page text for example. PatchesThis has been fixed in pypdf==6.14.1. WorkaroundsIf you...

Serena: Unauthenticated Flask dashboard on fixed port enables DNS rebinding → memory poisoning → RCE

SummarySerena's built-in web dashboard exposes an unauthenticated Flask API on a fixed, predictable port TCP 24282, hardcoded as 0x5EDA in constants.py. The server has no authentication, no CSRF protection, and no Host header validation. A DNS rebinding attack allows a malicious webpage to reach...

Kiwi TCMS's /init-db/ page renders and responds to requests after first use

Kiwi TCMS provides the /init-db/ page as part of its setup mechanism for administrators who prefer a browser instead of the command line. In previous versions of Kiwi TCMS this page still renders and responds to requests even after first use. ImpactThe /init-db/ page does not require any user...

Glances is Vulnerable to Command Injection via KVM/QEMU VM Domain Names in glances/plugins/vms/engines/virsh.py

SummaryThe Glances KVM/QEMU monitoring engine glances/plugins/vms/engines/virsh.py passes VM domain names, read directly from virsh list --all output, into f-string command templates that are processed by securepopen. securepopen is explicitly designed to interpret &&, |, and as shell operators...

Open WebUI: Forged chat-file link allows cross-user file read and deletion

SummaryOpen WebUI v0.9.5 lets an authenticated user attach arbitrary fileid values to their own chat message without checking whether they own or can read those files. If the attacker then shares that chat and grants themselves read access, hasaccesstofile treats the victim file as accessible...

vLLM: GGUF dequantize kernel int truncation exposes uninitialized GPU memory in multi-tenant serving

SummaryInteger truncation of tensor dimensions in vLLM's GGUF dequantize kernels csrc/quantization/gguf/ggufkernel.cu causes partial tensor processing. The output tensor is allocated at full size via torch::empty uninitialized memory, but the dequantize CUDA kernel processes only a truncated numb...

PraisonAI spider_tools SSRF protection bypass via alternate loopback host encodings

SummaryPraisonAI's spidertools URL validation can be bypassed using alternate loopback host encodings.The affected component is:textpraisonaiagents/tools/spidertools.pyThe tool contains a URL validation function intended to block local or unsafe targets before fetching attacker-controlled URLs...

Pydantic AI: SSRF cloud-metadata blocklist bypass via IPv4-mapped IPv6 (Incomplete fix of CVE-2026-25580)

SummaryWhen an application using Pydantic AI opts a URL into forcedownload='allow-local' which disables the default block on private/internal IPs, the cloud-metadata blocklist could be bypassed by encoding the metadata IP in an IPv6 transition form IPv4-mapped IPv6, 6to4, or NAT64. Dual-stack and...

Paramiko rsakey.py allows the SHA-1 algorithm

In Paramiko through 4.0.0 before a448945, rsakey.py allows the SHA-1 algorithm...

PraisonAI Has Unauthenticated SSE Event Stream that Exposes All Agent Activity in A2U Server

The A2U Agent-to-User event stream server in PraisonAI exposes all agent activity without authentication. This is a separate component from the gateway server fixed in CVE-2026-34952.The createa2uroutes function registers the following endpoints with NO authentication checks:- GET /a2u/info —...

PraisonAI has Memory State Leakage and Path Traversal in MultiAgent Context Handling

SummaryThe MultiAgentLedger and MultiAgentMonitor components in the provided code exhibit vulnerabilities that can lead to context leakage and arbitrary file operations. Specifically:1. Memory State Leakage via Agent ID Collision: The MultiAgentLedger uses a dictionary to store ledgers by agent I...

Langflow: Authenticated Users Can Read, Modify, and Delete Any Flow via Missing Ownership Check

Vulnerability IDOR in GET/PATCH/DELETE /api/v1/flow/flowidThe readflow helper in src/backend/base/langflow/api/v1/flows.py branched on the AUTOLOGIN setting to decide whether to filter by userid. When AUTOLOGIN was False i.e., authentication was enabled, neither branch enforced an ownership check...

python-ecdsa: Denial of Service via improper DER length validation in crafted private keys

SummaryAn issue in the low-level DER parsing functions can cause unexpected exceptions to be raised from the public API functions.1. ecdsa.der.removeoctetstring accepts truncated DER where the encoded length exceeds the available buffer. For example, an OCTET STRING that declares a length of 4096...

OpenStack Cinder, Glance, and Nova vulnerable to arbitrary file access

An issue was discovered in OpenStack Cinder through 24.0.0, Glance before 28.0.2, and Nova before 29.0.3. Arbitrary file access can occur via custom QCOW2 external data. By supplying a crafted QCOW2 image that references a specific data file path, an authenticated user may convince systems to...

Potential DoS via the Tudoor mechanism in eventlet and dnspython

eventlet before 0.35.2, as used in dnspython before 2.6.0, allows remote attackers to interfere with DNS name resolution by quickly sending an invalid packet from the expected IP address and source port, aka a "TuDoor" attack. In other words, dnspython does not have the preferred behavior in whic...

`MirrorPadGrad` heap out of bounds read

ImpactIf MirrorPadGrad is given outsize input paddings, TensorFlow will give a heap OOB error.pythonimport tensorflow as tftf.rawops.MirrorPadGradinput=1, paddings=0x77f00000,0xa000000, mode = 'REFLECT' PatchesWe have patched the issue in GitHub commit 717ca98d8c3bba348ff62281fdf38dcb5ea1ec92.The...

TensorFlow vulnerable to `CHECK` fail in `AvgPoolGrad`

ImpactThe implementation of AvgPoolGrad does not fully validate the input originputshape. This results in a CHECK failure which can be used to trigger a denial of service attack:pythonimport tensorflow as tfksize = 1, 2, 2, 1strides = 1, 2, 2, 1padding = "VALID"dataformat = "NHWC"originputshape =...

TensorFlow vulnerable to segfault in `LowerBound` and `UpperBound`

ImpactIf LowerBound or UpperBound is given an emptysortedinputs input, it results in a nullptr dereference, leading to a segfault that can be used to trigger a denial of service attack.pythonimport tensorflow as tfouttype = tf.int32sortedinputs = tf.constant, shape=10,0, dtype=tf.float32values =...

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,...

Out-of-bounds Read and Out-of-bounds Write in OpenCV

An issue was discovered in OpenCV before 3.4.7 and 4.x before 4.1.1 OpenCV-Python before 3.4.7.28 and 4.x before 4.1.1.26. There is an out of bounds read/write in the function HaarEvaluator::OptFeature::calc in modules/objdetect/src/cascadedetect.hpp, which leads to denial of service...

PYSEC-2026-3444

httplib2 is a comprehensive HTTP client library for Python. Prior to 0.32.0, httplib2 performs unbounded decompression of HTTP response bodies encoded with Content-Encoding: gzip or deflate in decompressContent in httplib2/init.py, allowing a malicious or compromised HTTP server to return a small...

PYSEC-2026-2323

Zeep is a Python SOAP client. From 4.0.0 before 4.3.3, Settings.forbidexternal is defined but not enforced when parsing WSDL or XSD documents, allowing transitive xsd:import, xsd:include, wsdl:import, and lxml entity or DTD references to fetch attacker-chosen HTTP or HTTPS URLs. This issue is fix...

Boltz contains an insecure deserialization vulnerability in its molecule loading functionality

Boltz 2.0.0 contains an insecure deserialization vulnerability in its molecule loading functionality. The application uses Python pickle to deserialize molecule data files without validation. An attacker with the ability to place a malicious pickle file in a directory processed by boltz can achie...

Malicious code in mem8 (PyPI)

Part of the "Hades" wave of the Shai-Hulud supply-chain campaign. On 2026-06-08,malicious phantom releases of mem8 were published to PyPI using stolencredentials. The package executes a bundled JavaScript payload via the Bunruntime on import that harvests and exfiltrates credentials and...

OctoPrint has Timing Side-Channel Vulnerability in API Key Authentication

ImpactOctoPrint versions up to and including 1.11.5 are affected by a theoretical timing attack vulnerability that allows API key extraction over the network.Due to using character based comparison that short-circuits on the first mismatched character during API key validation, rather than a...

FastAPI Api Key has a timing side-channel in verify_key that allows statistical key validity detection

ImpactTiming side-channel vulnerability in verifykey. The method applied a random delay only on verification failures, allowing an attacker to statistically distinguish valid from invalid API keys by measuring response latencies. With enough repeated requests, an adversary could infer whether a...

virtualenv Has TOCTOU Vulnerabilities in Directory Creation

ImpactTOCTOU Time-of-Check-Time-of-Use vulnerabilities in virtualenv allow local attackers to perform symlink-based attacks on directory creation operations. An attacker with local access can exploit a race condition between directory existence checks and creation to redirect virtualenv's appdata...

Werkzeug safe_join() allows Windows special device names with compound extensions

Werkzeug's safejoin function allows path segments with Windows device names that have file extensions or trailing spaces. On Windows, there are special device names such as CON, AUX, etc that are implicitly present and readable in every directory. Windows still accepts them with any file extensio...

pretix has Broken Access Control Allowing Cross-User File Access via UUID

An API endpoint allowed access to sensitive files from other users by knowing the UUID of the file that were not intended to be accessible by UUID only...

Ansible Community General Collection is vulnerable to exposure of sensitive information

A flaw was found in ansible-collection-community-general. This vulnerability allows for information exposure IE of sensitive credentials, specifically plaintext passwords, via verbose output when running Ansible with debug modes. Attackers with access to logs could retrieve these secrets and...

Apache Airflow has a command injection vulnerability in "example_dag_decorator"

An example dag exampledagdecorator had non-validated parameter that allowed the UI user to redirect the example to a malicious server and execute code on worker. This however required that the example dags are enabled in production not default or the example dag code copied to build your own...

Open WebUI Affected by an External Model Server (Direct Connections) Code Injection via SSE Events

SummaryOpen WebUI v0.6.33 and below contains a code injection vulnerability in the Direct Connections feature that allows malicious external model servers to execute arbitrary JavaScript in victim browsers via Server-Sent Event SSE execute events. This leads to authentication token theft, complet...

Authlib is vulnerable to Denial of Service via Oversized JOSE Segments

SummaryAuthlib’s JOSE implementation accepts unbounded JWS/JWT header and signature segments. A remote attacker can craft a token whose base64url‑encoded header or signature spans hundreds of megabytes. During verification, Authlib decodes and parses the full input before it is rejected, driving...

Fides Webserver API Rate Limiting Vulnerability in Proxied Environments

SummaryThe Fides Webserver API's built-in IP-based rate limiting is ineffective in environments with CDNs, proxies or load balancers. The system incorrectly applies rate limits based on directly connected infrastructure IPs rather than client IPs, and stores counters in-memory rather than in a...

Langroid has a Code Injection vulnerability in LanceDocChatAgent through vector_store

SummaryLanceDocChatAgent uses pandas eval through computefromdocs:https://github.com/langroid/langroid/blob/18667ec7e971efc242505196f6518eb19a0abc1c/langroid/vectorstore/base.pyL136-L150As a result, an attacker may be able to make the agent run malicious commands through QueryPlan.dataframecalc...

Transformers Regular Expression Denial of Service (ReDoS) vulnerability

A Regular Expression Denial of Service ReDoS vulnerability was identified in the huggingface/transformers library, specifically in the file tokenizationgptneoxjapanese.py of the GPT-NeoX-Japanese model. The vulnerability occurs in the SubWordJapaneseTokenizer class, where regular expressions...

Whoogle allows attackers to execute arbitrary code via supplying a crafted search query

An issue in the component /models/config.py of Whoogle search v0.9.0 allows attackers to execute arbitrary code via supplying a crafted search query...

Composio Eval Injection Vulnerability

In composiohq/composio version 0.4.3, the mathematicalcalculator endpoint uses the unsafe eval function to perform mathematical operations. This can lead to arbitrary code execution if untrusted input is passed to the eval function...

Prefect CORS (Cross-Origin Resource Sharing) misconfiguration

A CORS Cross-Origin Resource Sharing misconfiguration in prefecthq/prefect prior to version 3.0.3 allows unauthorized domains to access sensitive data. This vulnerability can lead to unauthorized access to the database, resulting in potential data leaks, loss of confidentiality, service disruptio...

FastChat Denial of Service vulnerability

A Denial of Service DoS vulnerability exists in the file upload feature of lm-sys/fastchat version 0.2.36. The vulnerability is due to improper handling of form-data with a large filename in the file upload request. An attacker can exploit this by sending a payload with an excessively large...

Feast Cross-Origin Resource Sharing vulnerability

A Cross-Origin Resource Sharing CORS vulnerability exists in feast-dev/feast version 0.40.0. The CORS configuration on the agentscope server does not properly restrict access to only trusted origins, allowing any external domain to make requests to the API. This can bypass intended security...

Gradio Vulnerable to Denial of Service (DoS) via Crafted HTTP Request

A Regular Expression Denial of Service ReDoS vulnerability exists in the gradio-app/gradio repository, affecting the gr.Datetime component. The affected version is git commit 98cbcae. The vulnerability arises from the use of a regular expression ^?:\snow\s?:-\s\d+\sdmhs??\s$ to process user input...

Letta (previously MemGPT) incorrect access control vulnerability

Incorrect access control in the /users endpoint of Cpacker MemGPT v0.3.17 allows attackers to access sensitive data...

Werkzeug possible resource exhaustion when parsing file data in forms

Applications using Werkzeug to parse multipart/form-data requests are vulnerable to resource exhaustion. A specially crafted form body can bypass the Request.maxformmemorysize setting.The Request.maxcontentlength setting, as well as resource limits provided by deployment software and platforms, a...

open-webui Insecure Direct Object Reference (IDOR) vulnerability

An Insecure Direct Object Reference IDOR vulnerability exists in open-webui/open-webui version v0.3.8. The vulnerability occurs in the API endpoint http://0.0.0.0:3000/api/v1/memories/id/update, where the decentralization design is flawed, allowing attackers to edit other users' memories without...