CVE Database

Out of bounds read in Web Speech in Google Chrome prior to 146.0.7680.71 allowed a remote attacker to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)

AdGuard Home is a network-wide software for blocking ads and tracking. Prior to 0.107.73, an unauthenticated remote attacker can bypass all authentication in AdGuardHome by sending an HTTP/1.1 request that requests an upgrade to HTTP/2 cleartext (h2c). Once the upgrade is accepted, the resulting HTTP/2 connection is handled by the inner mux, which has no authentication middleware attached. All subsequent HTTP/2 requests on that connection are processed as fully authenticated, regardless of whether any credentials were provided. This vulnerability is fixed in 0.107.73.

2FAuth is a web app to manage Two-Factor Authentication (2FA) accounts and generate their security codes. Prior to 6.1.0, a blind SSRF vulnerability exists in 2FAuth that allows authenticated users to make arbitrary HTTP requests from the server to internal networks and cloud metadata endpoints. The image parameter in OTP URL is not properly validated for internal / private IP addresses before making HTTP requests. While the previous fix added response validation to ensure only valid images are stored but HTTP request is still made to arbitrary URLs before this validation occurs. This vulnerability is fixed in 6.1.0.

Winter is a free, open-source content management system (CMS) based on the Laravel PHP framework. Prior to 1.0.477, 1.1.12, and 1.2.12, Winter CMS allowed authenticated backend users to escalate their accounts level of access to the system by modifying the roles / permissions assigned to their account through specially crafted requests to the backend while logged in. To actively exploit this security issue, an attacker would need access to the Backend with a user account with any level of access. This vulnerability is fixed in 1.0.477, 1.1.12, and 1.2.12.

An issue pertaining to CWE-259: Use of Hard-coded Password was discovered in oslabs-beta ThermaKube master.

An issue pertaining to CWE-89: Improper Neutralization of Special Elements used in an SQL Command was discovered in benkeen generatedata 4.0.14.

Insecure Access Control in Contact Plan, E-Mail, SMS and Fax components in Asseco SEE Live 2.0 allows remote attackers to access and execute attachments via a computable URL.

Plunk is an open-source email platform built on top of AWS SES. Prior to 0.7.0, a Server-Side Request Forgery (SSRF) vulnerability existed in the SNS webhook handler. An unauthenticated attacker could send a crafted request that caused the server to make an arbitrary outbound HTTP GET request to any host accessible from the server. This vulnerability is fixed in 0.7.0.

xygeni-action is the GitHub Action for Xygeni Scanner. On March 3, 2026, an attacker with access to compromised credentials created a series of pull requests (#46, #47, #48) injecting obfuscated shell code into action.yml. The PRs were blocked by branch protection rules and never merged into the main branch. However, the attacker used the compromised GitHub App credentials to move the mutable v5 tag to point at the malicious commit (4bf1d4e19ad81a3e8d4063755ae0f482dd3baf12) from one of the unmerged PRs. This commit remained in the repository's git object store, and any workflow referencing @v5 would fetch and execute it. This is a supply chain compromise via tag poisoning. Any GitHub Actions workflow referencing xygeni/xygeni-action@v5 during the affected window (approximately March 3–10, 2026) executed a C2 implant that granted the attacker arbitrary command execution on the CI runner for up to 180 seconds per workflow run.

Himmelblau is an interoperability suite for Microsoft Azure Entra ID and Intune. From 3.0.0 to before 3.1.0, if Himmelblau is deployed without a configured tenant domain in himmelblau.conf, authentication is not tenant-scoped. In this mode, Himmelblau can accept authentication attempts for arbitrary Entra ID domains by dynamically registering providers at runtime. This behavior is intended for initial/local bootstrap scenarios, but it can create risk in remote authentication environments. This vulnerability is fixed in 3.1.0.

Black is the uncompromising Python code formatter. Black provides a GitHub action for formatting code. This action supports an option, use_pyproject: true, for reading the version of Black to use from the repository pyproject.toml. A malicious pull request could edit pyproject.toml to use a direct URL reference to a malicious repository. This could lead to arbitrary code execution in the context of the GitHub Action. Attackers could then gain access to secrets or permissions available in the context of the action. Version 26.3.0 fixes this vulnerability.

WeGIA is a web manager for charitable institutions. Prior to version 3.6.6, a critical SQL injection vulnerability exists in the WeGIA application. The remover_produto_ocultar.php script uses extract($_REQUEST) to populate local variables and then directly concatenates these variables into a SQL query executed via PDO::query. This allows an authenticated (or auth-bypassed) attacker to execute arbitrary SQL commands. This can be used to exfiltrate sensitive data from the database or, as demonstrated in this PoC, cause a time-based delay (denial of service). This vulnerability is fixed in 3.6.6.

Unity Catalog is an open, multi-modal Catalog for data and AI. In 0.4.0 and earlier, a critical authentication bypass vulnerability exists in the Unity Catalog token exchange endpoint (/api/1.0/unity-control/auth/tokens). The endpoint extracts the issuer (iss) claim from incoming JWTs and uses it to dynamically fetch the JWKS endpoint for signature validation without validating that the issuer is a trusted identity provider.

Frappe is a full-stack web application framework. Prior to 15.84.0 and 14.99.0, a specially crafted request made to a certain endpoint could result in SQL injection, allowing an attacker to extract information they wouldn't otherwise be able to. This vulnerability is fixed in 15.84.0 and 14.99.0.

Taskosaur is an open source project management platform with conversational AI for task execution in-app. In 1.0.0, the application does not properly validate or restrict the role parameter during the user registration process. An attacker can manually modify the request payload and assign themselves elevated privileges. Because the backend does not enforce role assignment restrictions or ignore client-supplied role parameters, the server accepts the manipulated value and creates the account with SUPER_ADMIN privileges. This allows any unauthenticated attacker to register a fully privileged administrative account.

SAPIDO RB-1732 V2.0.43 contains a remote command execution vulnerability that allows unauthenticated attackers to execute arbitrary system commands by submitting malicious input to the formSysCmd endpoint. Attackers can send POST requests with the sysCmd parameter containing shell commands to execute code on the device with router privileges.

FileThingie 2.5.7 contains an arbitrary file upload vulnerability that allows attackers to upload malicious files by sending ZIP archives through the ft2.php endpoint. Attackers can upload ZIP files containing PHP shells, use the unzip functionality to extract them into accessible directories, and execute arbitrary commands through the extracted PHP files.

NetGain EM Plus 10.1.68 contains a remote code execution vulnerability that allows unauthenticated attackers to execute arbitrary system commands by submitting malicious parameters to the script_test.jsp endpoint. Attackers can send POST requests with shell commands embedded in the 'content' parameter to execute code and retrieve command output.

Epross AVCON6 systems management platform contains an object-graph navigation language (OGNL) injection vulnerability that allows unauthenticated attackers to execute arbitrary commands by injecting malicious OGNL expressions. Attackers can send crafted requests to the login.action endpoint with OGNL payloads in the redirect parameter to instantiate ProcessBuilder objects and execute system commands with root privileges.

Cloud CLI (aka Claude Code UI) is a desktop and mobile UI for Claude Code, Cursor CLI, Codex, and Gemini-CLI. Prior to 1.25.0, OS Command Injection via WebSocket Shell. Both projectPath and initialCommand in server/index.js are taken directly from the WebSocket message payload and interpolated into a bash command string without any sanitization, enabling arbitrary OS command execution. A secondary injection vector exists via unsanitized sessionId. This vulnerability is fixed in 1.25.0.

Parse Server is an open source backend that can be deployed to any infrastructure that can run Node.js. Prior to 9.6.0-alpha.5 and 8.6.31, a SQL injection vulnerability exists in the PostgreSQL storage adapter when processing Increment operations on nested object fields using dot notation (e.g., stats.counter). The sub-key name is interpolated directly into SQL string literals without escaping. An attacker who can send write requests to the Parse Server REST API can inject arbitrary SQL via a crafted sub-key name containing single quotes, potentially executing commands or reading data from the database, bypassing CLPs and ACLs. Only Postgres deployments are affected. This vulnerability is fixed in 9.6.0-alpha.5 and 8.6.31.

Cloud CLI (aka Claude Code UI) is a desktop and mobile UI for Claude Code, Cursor CLI, Codex, and Gemini-CLI. Prior to 1.24.0, multiple Git-related API endpoints use execAsync() with string interpolation of user-controlled parameters (file, branch, message, commit), allowing authenticated attackers to execute arbitrary OS commands. This vulnerability is fixed in 1.24.0.

Parse Server is an open source backend that can be deployed to any infrastructure that can run Node.js. A SQL injection vulnerability exists in the PostgreSQL storage adapter when processing Increment operations on nested object fields using dot notation (e.g., stats.counter). The amount value is interpolated directly into the SQL query without parameterization or type validation. An attacker who can send write requests to the Parse Server REST API can inject arbitrary SQL subqueries to read any data from the database, bypassing CLPs and ACLs. MongoDB deployments are not affected. This vulnerability is fixed in 9.6.0-alpha.3 and 8.6.29.

Jellyfin is an open-source media system. The code-quality.yml GitHub Actions workflow in jellyfin/jellyfin-ios is vulnerable to arbitrary code execution via pull requests from forked repositories. Due to the workflow's elevated permissions (nearly all write permissions), this vulnerability enables full repository takeover of jellyfin/jellyfin-ios, exfiltration of highly privileged secrets, Apple App Store supply chain attack, GitHub Container Registry (ghcr.io) package poisoning, and full jellyfin organization compromise via cross-repository token usage. Note: This is not a code vulnerability, but a vulnerability in the GitHub Actions workflows. No new version is required for this GHSA and end users do not need to take any actions.

Parse Server is an open source backend that can be deployed to any infrastructure that can run Node.js. Prior to 9.6.0-alpha.2 and 8.6.28, an attacker can use a dot-notation field name in combination with the sort query parameter to inject SQL into the PostgreSQL database through an improper escaping of sub-field values in dot-notation queries. The vulnerability may also affect queries that use dot-notation field names with the distinct and where query parameters. This vulnerability only affects deployments using a PostgreSQL database. This vulnerability is fixed in 9.6.0-alpha.2 and 8.6.28.

An edgecase in SSO implementation in Neo4j Enterprise edition versions prior to version 2026.02 can lead to unauthorised access under the following conditions: If a neo4j admin configures two or more OIDC providers AND configures one or more of them to be an authorization provider AND configures one or more of them to be authentication-only, then those that are authentication-only will also provide authorization. This edgecase becomes a security problem only if the authentication-only provider contains groups which have higher privileges than provided by the intended (configured) authorization provider. When using multiple plugins for authentication and authorisation, prior to the fix the issue could lead to a plugin configured to provide only authentication or authorisation capabilities erroneously providing both capabilities.  We recommend upgrading to versions 2026.02 (or 5.26.22) where the issue is fixed.

An issue in Lantronix EDS3000PS v.3.1.0.0R2 allows an attacker to execute arbitrary code and obtain sensitive information via the ltrx_evo component

An issue was discovered in Lantronix EDS3000PS 3.1.0.0R2. The host parameter of the TFTP client in the Filesystem Browser page is not properly sanitized. This can be exploited to escape from the original command and execute an arbitrary one with root privileges.

An issue was discovered in Lantronix EDS3000PS 3.1.0.0R2. The authentication on management pages can be bypassed by appending a specific suffix to the URL and by sending an Authorization header that uses "admin" as the username.

An issue was discovered in Lantronix EDS5000 2.1.0.0R3. The HTTP RPC module executes a shell command to write logs when user's authantication fails. The username is directly concatenated with the command without any sanitization. This allow attackers to inject arbitrary OS commands into the username parameter. Injected commands are executed with root privileges.