User ID references at mentions in document comments were not correctly sanitized. Script code could be injected to a users session when working with a malicious document. Please deploy the provided updates and patch releases. User-defined content like comments and mentions are now filtered to avoid potentially malicious content. No publicly available exploits are known.