Spring AI's spring-ai-neo4j-store contains a Cypher injection vulnerability in Neo4jVectorFilterExpressionConverter. When a user-controlled string is passed as a filter expression key in Neo4jVectorFilterExpressionConverter of spring-ai-neo4j-store, doKey() embeds the key into a backtick-delimited Cypher property accessor (node.`metadata.`) after stripping only double quotes, without escaping embedded backticks.This issue affects Spring AI: from 1.0.0 before 1.0.5, from 1.1.0 before 1.1.4.
Spring AI's spring-ai-bedrock-converse contains a Server-Side Request Forgery (SSRF) vulnerability in BedrockProxyChatModel when processing multimodal messages that include user-supplied media URLs. Insufficient validation of those URLs allows an attacker to induce the server to issue HTTP requests to unintended internal or external destinations.
This issue affects Spring AI: from 1.0.0 before 1.0.5, from 1.1.0 before 1.1.4.
A security vulnerability has been detected in Shenzhen Ruiming Technology Streamax Crocus up to 1.3.44. Affected is an unknown function of the file /RemoteFormat.do of the component Endpoint. Such manipulation of the argument State leads to sql injection. It is possible to launch the attack remotely. The exploit has been disclosed publicly and may be used. The vendor was contacted early about this disclosure but did not respond in any way.
A security flaw has been discovered in code-projects Simple Laundry System 1.0. This affects an unknown function of the file /modstaffinfo.php of the component Parameter Handler. The manipulation of the argument userid results in sql injection. The attack may be performed from remote. The exploit has been released to the public and may be used for attacks.
A vulnerability was determined in Tenda AC5 15.03.06.47. The affected element is the function decodePwd of the file /goform/WizardHandle of the component POST Request Handler. Executing a manipulation of the argument WANT/WANS can lead to stack-based buffer overflow. The attack can be executed remotely. The exploit has been publicly disclosed and may be utilized.
MyTube is a self-hosted downloader and player for several video websites Prior to version 1.8.72, an unauthenticated attacker can lock out administrator and visitor accounts from password-based authentication by triggering failed login attempts. The application exposes three password verification endpoints, all of which are publicly accessible. All three endpoints share a single file-backed login attempt state stored in `login-attempts.json`. When any endpoint records a failed authentication attempt via `recordFailedAttempt()`, the shared login attempt state is updated, increasing the `failedAttempts` counter and adjusting the associated timestamps and cooldown values. Before verifying a password, each endpoint calls `canAttemptLogin()`. This function checks the shared JSON file to determine whether a cooldown period is active. If the cooldown has not expired, the request is rejected before the password is validated. Because the failed attempt counter and cooldown timer are globally shared, failed authentication attempts against any endpoint affect all other endpoints. An attacker can exploit this by repeatedly sending invalid authentication requests to any of these endpoints, incrementing the shared counter and waiting for the cooldown period between attempts. By doing so, the attacker can progressively increase the lockout duration until it reaches 24 hours, effectively preventing legitimate users from authenticating. Once the maximum lockout is reached, the attacker can maintain the denial of service indefinitely by waiting for the cooldown to expire and sending another failed attempt, which immediately triggers another 24-hour lockout if no successful login occurred in the meantime. Version 1.8.72 fixes the vulnerability.
BuildKit is a toolkit for converting source code to build artifacts in an efficient, expressive and repeatable manner. Prior to version 0.28.1, when using a custom BuildKit frontend, the frontend can craft an API message that causes files to be written outside of the BuildKit state directory for the execution context. The issue has been fixed in v0.28.1. The vulnerability requires using an untrusted BuildKit frontend set with `#syntax` or `--build-arg BUILDKIT_SYNTAX`. Using these options with a well-known frontend image like `docker/dockerfile` is not affected.
cpp-httplib is a C++11 single-file header-only cross platform HTTP/HTTPS library. Prior to 0.39.0, the cpp-httplib HTTP client forwards stored Basic Auth, Bearer Token, and Digest Auth credentials to arbitrary hosts when following cross-origin HTTP redirects (301/302/307/308). A malicious or compromised server can redirect the client to an attacker-controlled host, which then receives the plaintext credentials in the `Authorization` header. Version 0.39.0 fixes the issue.
BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Prior to 1.4.37, the `docker.system_packages` field in `bentofile.yaml` accepts arbitrary strings that are interpolated directly into Dockerfile `RUN` commands without sanitization. Since `system_packages` is semantically a list of OS package names (data), users do not expect values to be interpreted as shell commands. A malicious `bentofile.yaml` achieves arbitrary command execution during `bentoml containerize` / `docker build`. Version 1.4.37 fixes the issue.
MyTube is a self-hosted downloader and player for several video websites Prior to version 1.8.69, an authorization bypass in the `/api/settings/import-database` endpoint allows attackers with low-privilege credentials to upload and replace the application's SQLite database entirely, leading to a full compromise of the application. The bypass is relevant for other POST routes as well. Version 1.8.69 fixes the issue.
Metabase is an open source business intelligence and embedded analytics tool. In Metabase Enterprise prior to versions 1.54.22, 1.55.22, 1.56.22, 1.57.16, 1.58.10, and 1.59.4, authenticated admins on Metabase Enterprise Edition can achieve Remote Code Execution (RCE) and Arbitrary File Read via the `POST /api/ee/serialization/import` endpoint. A crafted serialization archive injects an `INIT` property into the H2 JDBC spec, which can execute arbitrary SQL during a database sync. We confirmed this was possible on Metabase Cloud. This only affects Metabase Enterprise. Metabase OSS lacks the affected codepaths. All versions of Metabase Enterprise that have serialization, which dates back to at least version 1.47, are affected. Metabase Enterprise versions 1.54.22, 1.55.22, 1.56.22, 1.57.16, 1.58.10, and 1.59.4 patch the issue. As a workaround, disable the serialization import endpoint in their Metabase instance to prevent access to the vulnerable codepaths.
OpenHands is software for AI-driven development. Starting in version 1.5.0, a Command Injection vulnerability exists in the `get_git_diff()` method at `openhands/runtime/utils/git_handler.py:134`. The `path` parameter from the `/api/conversations/{conversation_id}/git/diff` API endpoint is passed unsanitized to a shell command, allowing authenticated attackers to execute arbitrary commands in the agent sandbox. The user is already allowed to instruct the agent to execute commands, but this bypasses the normal channels. Version 1.5.0 fixes the issue.
pypdf is a free and open-source pure-python PDF library. Versions prior to 6.9.2 have a vulnerability in which an attacker can craft a PDF which leads to an infinite loop. This requires reading a file in non-strict mode. This has been fixed in pypdf 6.9.2. If users cannot upgrade yet, consider applying the changes from the patch manually.
A vulnerability was found in Tenda AC5 15.03.06.47. Impacted is the function formWifiWpsOOB of the file /goform/WifiWpsOOB of the component POST Request Handler. Performing a manipulation of the argument index results in stack-based buffer overflow. Remote exploitation of the attack is possible. The exploit has been made public and could be used.
A vulnerability has been found in Tenda AC5 15.03.06.47. This issue affects the function formSetCfm of the file /goform/setcfm of the component POST Request Handler. Such manipulation of the argument funcpara1 leads to stack-based buffer overflow. The attack may be launched remotely. The exploit has been disclosed to the public and may be used.
Incus is a system container and virtual machine manager. Prior to version 6.23.0, the web server spawned by `incus webui` incorrectly validates the authentication token such that an invalid value will be accepted. `incus webui` runs a local web server on a random localhost port. For authentication, it provides the user with a URL containing an authentication token. When accessed with that token, Incus creates a cookie persisting that token without needing to include it in subsequent HTTP requests. While the Incus client correctly validates the value of the cookie, it does not correctly validate the token when passed int the URL.
This allows for an attacker able to locate and talk to the temporary web server on localhost to have as much access to Incus as the user who ran `incus webui`. This can lead to privilege escalation by another local user or an access to the user's Incus instances and possibly system resources by a remote attack able to trick the local user into interacting with the Incus UI web server. Version 6.23.0 patches the issue.
Cocos AI is a confidential computing system for AI. The current implementation of attested TLS (aTLS) in CoCoS is vulnerable to a relay attack affecting all versions from v0.4.0 through v0.8.2. This vulnerability is present in both the AMD SEV-SNP and Intel TDX deployment targets supported by CoCoS. In the affected design, an attacker may be able to extract the ephemeral TLS private key used during the intra-handshake attestation. Because the attestation evidence is bound to the ephemeral key but not to the TLS channel, possession of that key is sufficient to relay or divert the attested TLS session. A client will accept the connection under false assumptions about the endpoint it is communicating with — the attestation report cannot distinguish the genuine attested service from the attacker's relay. This undermines the intended authentication guarantees of attested TLS. A successful attack may allow an attacker to impersonate an attested CoCoS service and access data or operations that the client intended to send only to the genuine attested endpoint. Exploitation requires the attacker to first extract the ephemeral TLS private key, which is possible through physical access to the server hardware, transient execution attacks, or side-channel attacks. Note that the aTLS implementation was fully redesigned in v0.7.0, but the redesign does not address this vulnerability. The relay attack weakness is architectural and affects all releases in the v0.4.0–v0.8.2 range. This vulnerability class was formally analyzed and demonstrated across multiple attested TLS implementations, including CoCoS, by researchers whose findings were disclosed to the IETF TLS Working Group. Formal verification was conducted using ProVerif. As of time of publication, there is no patch available. No complete workaround is available. The following hardening measures reduce but do not eliminate the risk: Keep TEE firmware and microcode up to date to reduce the key-extraction surface; define strict attestation policies that validate all available report fields, including firmware versions, TCB levels, and platform configuration registers; and/or enable mutual aTLS with CA-signed certificates where deployment architecture permits.
Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to version 0.8.6, any authenticated user can overwrite any file's content by ID through the `POST /api/v1/retrieval/process/files/batch` endpoint. The endpoint performs no ownership check, so a regular user with read access to a shared knowledge base can obtain file UUIDs via `GET /api/v1/knowledge/{id}/files` and then overwrite those files, escalating from read to write. The overwritten content is served to the LLM via RAG, meaning the attacker controls what the model tells other users. Version 0.8.6 patches the issue.
vLLM is an inference and serving engine for large language models (LLMs). Starting in version 0.10.1 and prior to version 0.18.0, two model implementation files hardcode `trust_remote_code=True` when loading sub-components, bypassing the user's explicit `--trust-remote-code=False` security opt-out. This enables remote code execution via malicious model repositories even when the user has explicitly disabled remote code trust. Version 0.18.0 patches the issue.
A flaw has been found in Tenda AC5 15.03.06.47. This vulnerability affects the function formQuickIndex of the file /goform/QuickIndex of the component POST Request Handler. This manipulation of the argument PPPOEPassword causes stack-based buffer overflow. The attack may be initiated remotely. The exploit has been published and may be used.
A vulnerability was detected in Tenda AC5 15.03.06.47. This affects the function fromAddressNat of the file /goform/addressNat of the component POST Request Handler. The manipulation of the argument page results in stack-based buffer overflow. The attack can be launched remotely. The exploit is now public and may be used.
In TigerVNC before 1.16.2, Image.cxx in x0vncserver allows other users to observe or manipulate the screen contents, or cause an application crash, because of incorrect permissions.
Incus is a system container and virtual machine manager. Incus provides an API to retrieve VM screenshots. That API relies on the use of a temporary file for QEMU to write the screenshot to which is then picked up and sent to the user prior to deletion. As versions prior to 6.23.0 use predictable paths under /tmp for this, an attacker with local access to the system can abuse this mechanism by creating their own symlinks ahead of time. On the vast majority of Linux systems, this will result in a "Permission denied" error when requesting a screenshot. That's because the Linux kernel has a security feature designed to block such attacks, `protected_symlinks`. On the rare systems with this purposefully disabled, it's then possible to trick Incus intro truncating and altering the mode and permissions of arbitrary files on the filesystem, leading to a potential denial of service or possible local privilege escalation. Version 6.23.0 fixes the issue.
A memory leak exists in the Grassroots DICOM library (GDCM). The bug occurs when parsing malformed DICOM files with non-standard VR types in file meta information. The vulnerability leads to vast memory allocations and resource depletion, triggering a denial-of-service condition. A maliciously crafted file can fill the heap in a single read operation without properly releasing it.
Sharp is a content management framework built for Laravel as a package. Versions prior to 9.20.0 contain a vulnerability in the file upload endpoint that allows authenticated users to bypass all file type restrictions. The upload endpoint within the `ApiFormUploadController` accepts a client-controlled `validation_rule` parameter. This parameter is directly passed into the Laravel validator without sufficient server-side enforcement. By intercepting the request and sending `validation_rule[]=file`, an attacker can completely bypass all MIME type and file extension restrictions. This issue has been addressed in version 9.20.0 by removing the client-controlled validation rules and strictly defining upload rules server-side. As a workaround, ensure that the storage disk used for Sharp uploads is strictly private. Under default configurations, an attacker cannot directly execute uploaded PHP files unless a public disk configuration is explicitly used.
Sharp is a content management framework built for Laravel as a package. Versions prior to 9.20.0 have a path traversal vulnerability in the FileUtil class. The application fails to sanitize file extensions properly, allowing path separators to be passed into the storage layer. In `src/Utils/FileUtil.php`, the `FileUtil::explodeExtension()` function extracts a file's extension by splitting the filename at the last dot. This issue has been patched in version 9.20.0 by properly sanitizing the extension using `pathinfo(PATHINFO_EXTENSION)` instead of `strrpos()`, alongside applying strict regex replacements to both the base name and the extension.
PrestaShop is an open source e-commerce web application. Versions prior to 8.2.5 and 9.1.0 are vulnerable to stored Cross-Site Scripting (stored XSS) vulnerabilities in the BO. An attacker who can inject data into the database, via limited back-office access or a previously existing vulnerability, can exploit unprotected variables in back-office templates. Versions 8.2.5 and 9.1.0 contain a fix. No known workarounds are available.
Picomatch is a glob matcher written JavaScript. Versions prior to 4.0.4, 3.0.2, and 2.3.2 are vulnerable to Regular Expression Denial of Service (ReDoS) when processing crafted extglob patterns. Certain patterns using extglob quantifiers such as `+()` and `*()`, especially when combined with overlapping alternatives or nested extglobs, are compiled into regular expressions that can exhibit catastrophic backtracking on non-matching input. Applications are impacted when they allow untrusted users to supply glob patterns that are passed to `picomatch` for compilation or matching. In those cases, an attacker can cause excessive CPU consumption and block the Node.js event loop, resulting in a denial of service. Applications that only use trusted, developer-controlled glob patterns are much less likely to be exposed in a security-relevant way. This issue is fixed in picomatch 4.0.4, 3.0.2 and 2.3.2. Users should upgrade to one of these versions or later, depending on their supported release line. If upgrading is not immediately possible, avoid passing untrusted glob patterns to `picomatch`. Possible mitigations include disabling extglob support for untrusted patterns by using `noextglob: true`, rejecting or sanitizing patterns containing nested extglobs or extglob quantifiers such as `+()` and `*()`, enforcing strict allowlists for accepted pattern syntax, running matching in an isolated worker or separate process with time and resource limits, and applying application-level request throttling and input validation for any endpoint that accepts glob patterns.
Kestra is an open-source, event-driven orchestration platform Versions up to and including 1.3.3 render user-supplied flow YAML metadata fields — description, inputs[].displayName, inputs[].description — through the Markdown.vue component instantiated with html: true. The resulting HTML is injected into the DOM via Vue's v-html without any sanitization. This allows a flow author to embed arbitrary JavaScript that executes in the browser of any user who views or interacts with the flow. This is distinct from GHSA-r36c-83hm-pc8j / CVE-2026-29082, which covers only FilePreview.vue rendering .md files from execution outputs. The present finding affects different components, different data sources, and requires significantly less user interaction (zero-click for input.displayName). As of time of publication, it is unclear if a patch is available.
Pay is an open-source payment SDK extension package for various Chinese payment services. Prior to version 3.7.20, the `verify_wechat_sign()` function in `src/Functions.php` unconditionally skips all signature verification when the PSR-7 request reports `localhost` as the host. An attacker can exploit this by sending a crafted HTTP request to the WeChat Pay callback endpoint with a `Host: localhost` header, bypassing the RSA signature check entirely. This allows forging fake WeChat Pay payment success notifications, potentially causing applications to mark orders as paid without actual payment. Version 3.7.20 fixes the issue.
A vulnerability in Grafana Tempo exposes the S3 SSE-C encryption key in plaintext through the /status/config endpoint, potentially allowing unauthorized users to obtain the key used to encrypt trace data stored in S3.
Thanks to william_goodfellow for reporting this vulnerability.
A flaw was found in Red Hat OpenShift AI (RHOAI) llama-stack-operator. This vulnerability allows unauthorized access to Llama Stack services deployed in other namespaces via direct network requests, because no NetworkPolicy restricts access to the llama-stack service endpoint. As a result, a user in one namespace can access another user’s Llama Stack instance and potentially view or manipulate sensitive data.
The vulnerability exists in the UPnP component of TL-WR841N v14, where improper input validation leads to an out-of-bounds read, potentially causing a crash of the UPnP service.
Successful exploitation can cause the UPnP service to crash, resulting in a Denial-of-Service condition.Â
This vulnerability affects TL-WR841N v14 < EN_0.9.1 4.19 Build 260303 Rel.42399n (V14_260303) and < US_0.9.1.4.19 Build 260312 Rel. 49108n (V14_0304).
Incorrect Authorization vulnerability in Drupal AI (Artificial Intelligence) allows Resource Injection.This issue affects AI (Artificial Intelligence): from 0.0.0 before 1.1.11, from 1.2.0 before 1.2.12.
Fireshare facilitates self-hosted media and link sharing. In version 1.5.1, an authenticated path traversal vulnerability in Fireshare’s chunked upload endpoint allows an attacker to write arbitrary files outside the intended upload directory. The `checkSum` multipart field is used directly in filesystem path construction without sanitization or containment checks. This enables unauthorized file writes to attacker-chosen paths writable by the Fireshare process (e.g., container `/tmp`), violating integrity and potentially enabling follow-on attacks depending on deployment. Version 1.5.2 fixes the issue.
PinchTab is a standalone HTTP server that gives AI agents direct control over a Chrome browser. PinchTab `v0.8.3` through `v0.8.5` allow arbitrary JavaScript execution through `POST /wait` and `POST /tabs/{id}/wait` when the request uses `fn` mode, even if `security.allowEvaluate` is disabled. `POST /evaluate` correctly enforces the `security.allowEvaluate` guard, which is disabled by default. However, in the affected releases, `POST /wait` accepted a user-controlled `fn` expression, embedded it directly into executable JavaScript, and evaluated it in the browser context without checking the same policy. This is a security-policy bypass rather than a separate authentication bypass. Exploitation still requires authenticated API access, but a caller with the server token can execute arbitrary JavaScript in a tab context even when the operator explicitly disabled JavaScript evaluation. The current worktree fixes this by applying the same policy boundary to `fn` mode in `/wait` that already exists on `/evaluate`, while preserving the non-code wait modes. As of time of publication, a patched version is not yet available.
ClearanceKit intercepts file-system access events on macOS and enforces per-process access policies. In versions on the 4.1 branch and earlier, the opfilter Endpoint Security system extension enforced file access policy exclusively by intercepting ES_EVENT_TYPE_AUTH_OPEN events. Seven additional file operation event types were not intercepted, allowing any locally running process to bypass the configured FAA policy without triggering a denial. Commit a3d1733 adds subscriptions for all seven event types and routes them through the existing FAA policy evaluator. AUTH_RENAME and AUTH_UNLINK additionally preserve XProtect change detection: events on the XProtect path are allowed and trigger the existing onXProtectChanged callback rather than being evaluated against user policy. All versions on the 4.2 branch contain the fix. No known workarounds are available.
InvenTree is an Open Source Inventory Management System. Prior to version 1.2.6, certain API endpoints associated with bulk data operations can be hijacked to exfiltrate sensitive information from the database. The bulk operation API endpoints (e.g. `/api/part/`, `/api/stock/`, `/api/order/so/allocation/`, and others) accept a filters parameter that is passed directly to Django's ORM queryset.filter(**filters) without any field allowlisting. This enables any authenticated user to traverse model relationships using Django's __ lookup syntax and perform blind boolean-based data extraction. This issue is patched in version 1.2.6, and 1.3.0 (or above). Users should update to the patched versions. No known workarounds are available.
Boolean XPath expressions that evaluate to true can cause an infinite loop in logicalQuery.Select, leading to 100% CPU usage. This can be triggered by top-level selectors such as "1=1" or "true()".
The DataRow.Decode function fails to properly validate field lengths. A malicious or compromised PostgreSQL server can send a DataRow message with a negative field length, causing a slice bounds out of range panic.
The Delete function fails to properly validate offsets when processing malformed JSON input. This can lead to a negative slice index and a runtime panic, allowing a denial of service attack.
The msgpack decoder fails to properly validate the input buffer length when processing truncated fixext data (format codes 0xd4-0xd8). This can lead to an out-of-bounds read and a runtime panic, allowing a denial of service attack.
Ruckus Unleashed contains a remote code execution vulnerability in the web-based management interface that allows authenticated remote attackers to execute arbitrary code on the system when gateway mode is enabled. Attackers can exploit this vulnerability by sending specially crafted requests through the management interface to achieve arbitrary code execution on affected systems.
Impact:
A bad regular expression is generated any time you have multiple sequential optional groups (curly brace syntax), such as `{a}{b}{c}:z`. The generated regex grows exponentially with the number of groups, causing denial of service.
Patches:
Fixed in version 8.4.0.
Workarounds:
Limit the number of sequential optional groups in route patterns. Avoid passing user-controlled input as route patterns.
Ory Polis, formerly known as BoxyHQ Jackson, bridges or proxies a SAML login flow to OAuth 2.0 or OpenID Connect. Versions prior to 26.2.0 contain a DOM-based Cross-Site Scripting (XSS) vulnerability in Ory Polis's login functionality. The application improperly trusts a URL parameter (`callbackUrl`), which is passed to `router.push`. An attacker can craft a malicious link that, when opened by an authenticated user (or an unauthenticated user that later logs in), performs a client-side redirect and executes arbitrary JavaScript in the context of their browser. This could lead to credential theft, internal network pivoting, and unauthorized actions performed on behalf of the victim. Version 26.2.0 contains a patch for the issue.
Ory Keto is am open source authorization server for managing permissions at scale. Prior to version 26.2.0, the GetRelationships API in Ory Keto is vulnerable to SQL injection due to flaws in its pagination implementation. Pagination tokens are encrypted using the secret configured in `secrets.pagination`. An attacker who knows this secret can craft their own tokens, including malicious tokens that lead to SQL injection. If this configuration value is not set, Keto falls back to a hard-coded default pagination encryption secret. Because this default value is publicly known, attackers can generate valid and malicious pagination tokens manually for installations where this secret is not set. This issue can be exploited when GetRelationships API is directly or indirectly accessible to the attacker, the attacker can pass a raw pagination token to the affected API, and the configuration value `secrets.pagination` is not set or known to the attacker. An attacker can execute arbitrary SQL queries through forged pagination tokens. As a first line of defense, immediately configure a custom value for `secrets.pagination` by generating a cryptographically secure random secret. Next, upgrade Keto to a fixed version, 26.2.0 or later, as soon as possible.
Zen C is a systems programming language that compiles to human-readable GNU C/C11. Prior to version 0.4.4, a stack-based buffer overflow vulnerability in the Zen C compiler allows attackers to cause a compiler crash or potentially execute arbitrary code by providing a specially crafted Zen C source file (`.zc`) with excessively long struct, function, or trait identifiers. Users are advised to update to Zen C version v0.4.4 or later to receive a patch.
Tandoor Recipes is an application for managing recipes, planning meals, and building shopping lists. Versions up to and including 2.5.3 set ALLOWED_HOSTS = '*' by default, which causes Django to accept any value in the HTTP Host header without validation. The application uses request.build_absolute_uri() to generate absolute URLs in multiple contexts, including invite link emails, API pagination, and OpenAPI schema generation. An attacker who can send requests to the application with a crafted Host header can manipulate all server-generated absolute URLs. The most critical impact is invite link poisoning: when an admin creates an invite and the application sends the invite email, the link points to the attacker's server instead of the real application. When the victim clicks the link, the invite token is sent to the attacker, who can then use it at the real application. As of time of publication, it is unknown if a patched version is available.