An issue was discovered in Mahara 23.04.8 and 24.04.4. The use of a malicious export download URL can allow an attacker to download files that they do not have permission to download.
Mahara before 22.10.6, 23.04.6, and 24.04.1 allows cross-site scripting (XSS) via a file, with JavaScript code as part of its name, that is uploaded via the Mahara filebrowser system.
Supported versions of Mahara 24.04 before 24.04.1 and 23.04 before 23.04.6 are vulnerable to information being disclosed to an institution administrator under certain conditions via the 'Current submissions' page: Administration -> Groups -> Submissions.
Mahara before 24.04.9 exposes database connection information if the database becomes unreachable, e.g., due to the database server being temporarily down or too busy.
An issue was discovered in Mahara 23.04.8 and 24.04.4. Attackers may utilize escalation of privileges in certain cases when logging into Mahara with Learning Tools Interoperability (LTI).
An issue was discovered in Mahara 24.04 before 24.04.2 and 23.04 before 23.04.7. The About, Contact, and Help footer links can be set up to be vulnerable to Cross Site Scripting (XSS) due to not sanitising the values. These links can only be set up by an admin but are clickable by any logged-in person.
Mahara before 22.10.4 and 23.x before 23.04.4 allows information disclosure if the experimental HTML bulk export is used via the administration interface or via the CLI, and the resulting export files are given to the account holders. They may contain images of other account holders because the cache is not cleared after the files of one account are exported.
Mahara 21.10 before 21.10.6, 22.04 before 22.04.4, and 22.10 before 22.10.1 allows unsafe font upload for skins. A particularly structured XML file could allow one to traverse the server to obtain access to secure files or cause code execution based on the payload.
Mahara 21.10 before 21.10.6, 22.04 before 22.04.4, and 22.10 before 22.10.1 deserializes user input unsafely during skin import. A particularly structured XML file could cause code execution when being processed.
Mahara 21.04 before 21.04.7, 21.10 before 21.10.5, 22.04 before 22.04.3, and 22.10 before 22.10.0 potentially allow a PDF export to trigger a remote shell if the site is running on Ubuntu and the flag -dSAFER is not set with Ghostscript.
In Mahara 21.04 before 21.04.7, 21.10 before 21.10.5, 22.04 before 22.04.3, and 22.10 before 22.10.0, embedded images are accessible without a sufficient permission check under certain conditions.
In Mahara before 20.10.5, 21.04.4, 21.10.2, and 22.04.0, a site using Isolated Institutions is vulnerable if more than ten groups are used. They are all shown from page 2 of the group results list (rather than only being shown for the institution that the viewer is a member of).
Mahara before 20.10.5, 21.04.4, 21.10.2, and 22.04.0 allows stored XSS when a particular Cascading Style Sheets (CSS) class for embedly is used, and JavaScript code is constructed to perform an action.
Mahara before 20.10.5, 21.04.4, 21.10.2, and 22.04.0 is vulnerable to Cross Site Request Forgery (CSRF) because randomly generated tokens are too easily guessable.
In Mahara 21.04 before 21.04.3 and 21.10 before 21.10.1, portfolios created in groups that have not been shared with non-group members and portfolios created on the site and institution levels can be viewed without requiring a login if the URL to these portfolios is known.
In Mahara 20.10 before 20.10.4, 21.04 before 21.04.3, and 21.10 before 21.10.1, the names of folders in the Files area can be seen by a person not owning the folders. (Only folder names are affected. Neither file names nor file contents are affected.)
In Mahara before 20.04.5, 20.10.3, 21.04.2, and 21.10.0, the account associated with a web services token is vulnerable to being exploited and logged into, resulting in information disclosure (at a minimum) and often escalation of privileges.
In Mahara before 20.04.5, 20.10.3, 21.04.2, and 21.10.0, exported CSV files could contain characters that a spreadsheet program could interpret as a command, leading to execution of a malicious string locally on a device, aka CSV injection.