Is there a way to re-run only the failed prompts in the eval? I tried

--filter-failing

but it seems broken, see [#5755](https://github.com/promptfoo/promptfoo/issues/5755). Is there a workaround?

Hey! I have a fix for this issue here: https://github.com/promptfoo/promptfoo/pull/5770. This should be included in the next release