Hey, all –
Thanks for raising this issue. I'll address the headers individually.

Strict-Transport-Security – This is in place, as the report from securityheaders.com shows.
Content-Security-Policy – This is something we're currently working on. Implementing CSP is a complex undertaking as any misconfiguration can impact application functionality, but it is in progress. In the meantime we use other security headers, same-origin policies, input validation and HTML output encoding to mitigate many risks that can also be addressed by CSP.
X-Content-Type-Options – This header is present on endpoints where it provides a specific security benefit, such as attachment responses where mime sniffing attacks can occur. It's not implemented on all Zendesk endpoints, however.
Referrer-Policy – We are aware of this, but have no plans to implement it at this time.

For those who are looking for more in this area, we'd love to hear specifics.