Nekros1Xx

**Name of the Vulnerable Software and Affected Versions** Langflow versions prior to 1.9.0 **Description** Langflow contains a path traversal flaw in the Knowledge Bases API endpoint 'POST /api/v1/knowledge bases'. The issue resides in the `create knowledge base()` function, where the `name` variable is used to construct file paths without proper sanitization or containment checks. An authenticated attacker can provide traversal sequences or absolute paths in the `name` field to force the application to create directories and write files, specifically `embedding metadata.json` and `schema.json`, at any reachable location on the server's filesystem where the application has write permissions. This can lead to cross-user data compromise, arbitrary filesystem manipulation, and the corruption of existing knowledge bases through data overwrite. **Recommendations** Update to version 1.9.0.

**Name of the Vulnerable Software and Affected Versions** kafka-sink-azure-kusto versions prior to 5.2.3 **Description** The plugin fails to sanitize user-controlled values within the `kusto.tables.topics.mapping` configuration. Specifically, the `db`, `table`, `mapping`, and `format` fields are interpolated directly into Kusto Query Language (KQL) management and query commands using the `String.formatted()` function. An attacker with permissions to submit or edit Kafka Connect connector configurations can embed KQL metacharacters such as `;`, `|`, and `'` to execute arbitrary management commands. This allows for schema enumeration or modification, tampering with ingestion mappings, and altering streaming or retention policies on the target Azure Data Explorer database using the connector's service principal privileges. **Recommendations** Update kafka-sink-azure-kusto to version 5.2.3.

**Name of the Vulnerable Software and Affected Versions** Langflow versions prior to 1.9.0 **Description** Langflow is a tool for building and deploying AI-powered agents and workflows. A path traversal flaw exists in the Knowledge Bases API endpoint "DELETE /api/v1/knowledge bases" within the `delete knowledge bases bulk()` function. The issue arises because user-supplied knowledge base names provided in the `kb names` parameter are concatenated directly into file paths without proper sanitization or boundary validation. This allows an authenticated attacker to use traversal sequences (e.g., `../`) to escape the intended directory and trigger the `shutil.rmtree()` function to delete arbitrary directories on the server's filesystem. This can lead to cross-tenant data loss, deletion of critical application files, and potential service disruption. **Recommendations** Update to version 1.9.0. As a temporary workaround, restrict access to the "DELETE /api/v1/knowledge bases" API endpoint to minimize the risk of exploitation.

**Name of the Vulnerable Software and Affected Versions** Beszel versions prior to 0.18.2 Beszel versions 0.18.2 through 0.18.3 **Description** Beszel is a server monitoring platform. The platform’s authenticated API endpoints, specifically ''/api/beszel/containers/logs'' and ''/api/beszel/containers/info'', pass the `container` query parameter to the agent without proper validation. The agent then uses this parameter to construct Docker Engine API URLs using `fmt.Sprintf` instead of `url.PathEscape`. Because Go’s `http.Client` does not sanitize `../` sequences in URL paths sent over unix sockets, an authenticated user, even with a readonly role, can potentially traverse to arbitrary Docker API endpoints on the agent hosts. This could expose sensitive infrastructure details. **Recommendations** Update Beszel to version 0.18.4 or later. Update Beszel to version 0.18.4 or later.

**Name of the Vulnerable Software and Affected Versions** WPGraphQL versions prior to 2.9.1 **Description** The WPGraphQL software includes a GraphQL API for WordPress sites. A GitHub Actions workflow file (`release.yml`) in the `wp-graphql/wp-graphql` repository is susceptible to OS command injection. This occurs through the direct use of `${{ github.event.pull request.body }}` within a `run:` shell block. When a pull request is merged from the `develop` branch to the `master` branch, the content of the pull request body is directly inserted into a shell command, potentially enabling arbitrary command execution on the Actions runner. **Recommendations** Update to version 2.9.1 or later.

**Name of the Vulnerable Software and Affected Versions** LiveCode versions prior to commit e151c64c2bd80d2d53ac1333f1df9429fe6a1a11 **Description** LiveCode is an open-source, client-side code playground. The `i18n-update-pull` GitHub Actions workflow is susceptible to JavaScript injection prior to commit e151c64c2bd80d2d53ac1333f1df9429fe6a1a11. The title of a Pull Request is directly interpolated into a JavaScript block within the `actions/github-script` action using a GitHub Actions template expression. An attacker can inject arbitrary JavaScript by creating a Pull Request with a specially crafted title. This injected JavaScript executes with the permissions of the CI bot token (`CI APP ID` / `CI APP PRIVATE KEY`), potentially allowing for the exfiltration of repository secrets and unauthorized operations via the GitHub API. **Recommendations** Update LiveCode to commit e151c64c2bd80d2d53ac1333f1df9429fe6a1a11 or later.