So far I've only seen it in https://github.com/Sjors/sv2-tp/pull/10: - https://github.com/Sjors/sv2-tp/actions/runs/17395389917/job/49376250892?pr=10 - https://github.com/Sjors/sv2-tp/actions/runs/17395389917/job/49376250881?pr=10 It happened after I rebased on #12 (cc @Eunovo), but this may be a coincidence. This test has been flaky in the past.