Deleting users with certain names caused system files to be deleted. Risk is higher for systems which allow users to register themselves and have the data directory in the web root. This affects ownCloud/core versions < 10.6.
The fetch function in OAuth/Curl.php in Dropbox-PHP, as used in ownCloud Server before 6.0.8, 7.x before 7.0.6, and 8.x before 8.0.4 when an external Dropbox storage has been mounted, allows remote administrators of Dropbox.com to read arbitrary files via an @ (at sign) character in unspecified POST values.
Cross-site scripting (XSS) vulnerability in ownCloud before 6.0.1 allows remote authenticated users to inject arbitrary web script or HTML via the filename of an uploaded file.
ownCloud Server before 8.2.12, 9.0.x before 9.0.10, 9.1.x before 9.1.6, and 10.0.x before 10.0.2 are vulnerable to XSS on error pages by injecting code in url parameters.
Inadequate escaping lead to XSS vulnerability in the search module in ownCloud Server before 8.2.12, 9.0.x before 9.0.10, 9.1.x before 9.1.6, and 10.0.x before 10.0.2. To be exploitable a user has to write or paste malicious content into the search dialogue.
A logical error in ownCloud Server before 10.0.2 caused disclosure of valid share tokens for public calendars. Thus granting an attacker potentially access to publicly shared calendars without knowing the share token.
Nextcloud Server before 9.0.52 & ownCloud Server before 9.0.4 are vulnerable to a log pollution vulnerability potentially leading to a local XSS. The download log functionality in the admin screen is delivering the log in JSON format to the end-user. The file was delivered with an attachment disposition forcing the browser to download the document. However, Firefox running on Microsoft Windows would offer the user to open the data in the browser as an HTML document. Thus any injected data in the log would be executed.
Nextcloud Server before 9.0.52 & ownCloud Server before 9.0.4 are vulnerable to a content-spoofing attack in the files app. The location bar in the files app was not verifying the passed parameters. An attacker could craft an invalid link to a fake directory structure and use this to display an attacker-controlled error message to the user.