In Zulip Server versions from 1.7.0 to before 2.0.7, a bug in the new user signup process meant that users who registered their account using social authentication (e.g., GitHub or Google SSO) in an organization that also allows password authentication could have their personal API key stolen by an unprivileged attacker, allowing nearly full access to the user's account.

Zulip server before 2.0.5 incompletely validated the MIME types of uploaded files. A user who is logged into the server could upload files of certain types to mount a stored cross-site scripting attack on other logged-in users. On a Zulip server using the default local uploads backend, the attack is only effective against browsers lacking support for Content-Security-Policy such as Internet Explorer 11. On a Zulip server using the S3 uploads backend, the attack is confined to the origin of the configured S3 uploads hostname and cannot reach the Zulip server itself.

The Markdown parser in Zulip server before 2.0.5 used a regular expression vulnerable to exponential backtracking. A user who is logged into the server could send a crafted message causing the server to spend an effectively arbitrary amount of CPU time and stall the processing of future messages.

In Zulip Server versions before 1.7.2, there was an XSS issue with user uploads and the (default) LOCAL_UPLOADS_DIR storage backend.

In Zulip Server versions before 1.7.2, there was an XSS issue with stream names in topic typeahead.

In Zulip Server versions 1.5.x, 1.6.x, and 1.7.x before 1.7.2, there was an XSS issue with muting notifications.

In Zulip Server versions before 1.7.2, there were XSS issues with the frontend markdown processor.

In zulip before 1.3.12, deactivated users could access messages if SSO was enabled.

In zulip before 1.3.12, bot API keys were accessible to other users in the same realm.