Security Vulnerabilities (Directory traversal)

TorchServe is a tool for serving and scaling PyTorch models in production. Starting in version 0.1.0 and prior to version 0.9.0, using the model/workflow management API, there is a chance of uploading potentially harmful archives that contain files that are extracted to any location on the filesystem that is within the process permissions. Leveraging this issue could aid third-party actors in hiding harmful code in open-source/public models, which can be downloaded from the internet, and take advantage of machines running Torchserve. The ZipSlip issue in TorchServe has been fixed by validating the paths of files contained within a zip archive before extracting them. TorchServe release 0.9.0 includes fixes to address the ZipSlip vulnerability.

Directory Traversal vulnerability in TerraMaster v.s1.0 through v.2.295 allows a remote attacker to obtain sensitive information via a crafted GET request.

A CWE-23: Relative Path Traversal vulnerability exists in Telit Cinterion BGS5, Telit Cinterion EHS5/6/8, Telit Cinterion PDS5/6/8, Telit Cinterion ELS61/81, Telit Cinterion PLS62 that could allow a local, low privileged attacker to escape from virtual directories and get read/write access to protected files on the targeted system.

Directory Traversal vulnerability in jeecg-boot v.3.6.0 allows a remote privileged attacker to obtain sensitive information via the file directory structure.

Headwind MDM Web panel 5.22.1 is vulnerable to Directory Traversal.

Directory traversal vulnerability in CubeCart prior to 6.5.3 allows a remote authenticated attacker with an administrative privilege to obtain files in the system.

In mprivacy-tools before 2.0.406g in m-privacy TightGate-Pro Server, a Directory Traversal in the print function of the VNC service allows authenticated attackers (with access to a VNC session) to automatically transfer malicious PDF documents by moving them into the .spool directory, and then sending a signal to the VNC service, which automatically transfers them to the connected VNC client's filesystem.

In SysAid On-Premise before 23.3.36, a path traversal vulnerability leads to code execution after an attacker writes a file to the Tomcat webroot, as exploited in the wild in November 2023.

Dreamer CMS before version 4.0.1 is vulnerable to Directory Traversal. Background template management allows arbitrary modification of the template file, allowing system sensitive files to be read.

Peppermint Ticket Management through 0.2.4 allows remote attackers to read arbitrary files via a /api/v1/ticket/1/file/download?filepath=../ POST request.

Peppermint Ticket Management before 0.2.4 allows remote attackers to read arbitrary files via a /api/v1/users/file/download?filepath=./../ POST request.

In the module "Product Catalog (CSV, Excel, XML) Export PRO" (exportproducts) in versions up to 4.1.1 from MyPrestaModules for PrestaShop, a guest can download personal information without restriction by performing a path traversal attack. Due to a lack of permissions control and a lack of control in the path name construction, a guest can perform a path traversal to view all files on the information system.

Squidex is an open source headless CMS and content management hub. Affected versions are subject to an arbitrary file write vulnerability in the backup restore feature which allows an authenticated attacker to gain remote code execution (RCE). Squidex allows users with the `squidex.admin.restore` permission to create and restore backups. Part of these backups are the assets uploaded to an App. For each asset, the backup zip archive contains a `.asset` file with the actual content of the asset as well as a related `AssetCreatedEventV2` event, which is stored in a JSON file. Amongst other things, the JSON file contains the event type (`AssetCreatedEventV2`), the ID of the asset (`46c05041-9588-4179-b5eb-ddfcd9463e1e`), its filename (`test.txt`), and its file version (`0`). When a backup with this event is restored, the `BackupAssets.ReadAssetAsync` method is responsible for re-creating the asset. For this purpose, it determines the name of the `.asset` file in the zip archive, reads its content, and stores the content in the filestore. When the asset is stored in the filestore via the UploadAsync method, the assetId and fileVersion are passed as arguments. These are further passed to the method GetFileName, which determines the filename where the asset should be stored. The assetId is inserted into the filename without any sanitization and an attacker with squidex.admin.restore privileges to run arbitrary operating system commands on the underlying server (RCE).

FOG is a free open-source cloning/imaging/rescue suite/inventory management system. Prior to version 1.5.10, an endpoint intended to offer limited enumeration abilities to authenticated users was accessible to unauthenticated users. This enabled unauthenticated users to discover files and their respective paths that were visible to the Apache user group. Version 1.5.10 contains a patch for this issue.

sbt is a build tool for Scala, Java, and others. Given a specially crafted zip or JAR file, `IO.unzip` allows writing of arbitrary file. This would have potential to overwrite `/root/.ssh/authorized_keys`. Within sbt's main code, `IO.unzip` is used in `pullRemoteCache` task and `Resolvers.remote`; however many projects use `IO.unzip(...)` directly to implement custom tasks. This vulnerability has been patched in version 1.9.7.

Parse Server is an open source backend that can be deployed to any infrastructure that can run Node.js. Parse Server crashes when uploading a file without extension. This vulnerability has been patched in versions 5.5.6 and 6.3.1.

GibbonEdu Gibbon through version 25.0.0 allows Directory Traversal via the report template builder. An attacker can create a new Asset Component. The templateFileDestination parameter can be set to an arbitrary pathname (and extension). This allows creation of PHP files outside of the uploads directory, directly in the webroot.

The Learning Module in ILIAS 7.25 (2023-09-12 release) allows an attacker (with basic user privileges) to achieve a high-impact Directory Traversal attack on confidentiality and availability. By exploiting this network-based vulnerability, the attacker can move specified directories, normally outside the documentRoot, to a publicly accessible location via the PHP function rename(). This results in a total loss of confidentiality, exposing sensitive resources, and potentially denying access to the affected component and the operating system's components. To exploit this, an attacker must manipulate a POST request during the creation of an exercise unit, by modifying the old_name and new_name parameters via directory traversal. However, it's essential to note that, when exploiting this vulnerability, the specified directory will be relocated from its original location, rendering all files obtained from there unavailable.

ILIAS (2013-09-12 release) contains a medium-criticality Directory Traversal local file inclusion vulnerability in the ScormAicc module. An attacker with a privileged account, typically holding the tutor role, can exploit this to gain unauthorized access to and potentially retrieve confidential files stored on the web server. The attacker can access files that are readable by the web server user www-data; this may include sensitive configuration files and documents located outside the documentRoot. The vulnerability is exploited by an attacker who manipulates the file parameter in a URL, inserting directory traversal sequences in order to access unauthorized files. This manipulation allows the attacker to retrieve sensitive files, such as /etc/passwd, potentially compromising the system's security. This issue poses a significant risk to confidentiality and is remotely exploitable over the internet.

qdPM 9.2 allows Directory Traversal to list files and directories by navigating to the /uploads URI.

Artifact Hub is a web-based application that enables finding, installing, and publishing packages and configurations for CNCF projects. During a security audit of Artifact Hub's code base a security researcher identified a bug in which by using symbolic links in certain kinds of repositories loaded into Artifact Hub, it was possible to read internal files. Artifact Hub indexes content from a variety of sources, including git repositories. When processing git based repositories, Artifact Hub clones the repository and, depending on the artifact kind, reads some files from it. During this process, in some cases, no validation was done to check if the file was a symbolic link. This made possible to read arbitrary files in the system, potentially leaking sensitive information. This issue has been resolved in version `1.16.0`. Users are advised to upgrade. There are no known workarounds for this vulnerability.

Lack of sufficient path validation in South River Technologies' Titan MFT and Titan SFTP servers on Windows and Linux allows an authenticated attacker with administrative privileges to read any file on the filesystem via path traversal

Lack of sufficient path validation in South River Technologies' Titan MFT and Titan SFTP servers on Linux allows an authenticated attacker to get the size of an arbitrary file on the filesystem using path traversal in the ftp "SIZE" command

Insufficient path validation when writing a file via WebDAV in South River Technologies' Titan MFT and Titan SFTP servers on Linux allows an authenticated attacker to write a file to any location on the filesystem via path traversal

Insufficient path validation when extracting a zip archive in South River Technologies' Titan MFT and Titan SFTP servers on Windows and Linux allows an authenticated attacker to write a file to any location on the filesystem via path traversal