The two CSP frame-ancestor reporting tests fail in Firefox as they trigger a download there and the about:blank iframe after the download is not blocked. The other browsers do not trigger a download.

I don't know whether the response should trigger a download or not and if a frame-ancestor on the downloaded iframe should have any effect.
However, the test should test for the normal reporting behavior and not the download-edge case behavior. Simply, adding a content-type: text/html header to the response fixes the incorrect results in Firefox while not changing the results in the other browsers.

Tests: https://wpt.fyi/results/content-security-policy/reporting?label=master&label=experimental&aligned&q=frame-ancestors
WPT.fyi: https://wpt.live/content-security-policy/reporting/report-frame-ancestors.sub.html (triggers download)
Helper function that should be changed (add content-type):