And one more tiny comment: "the log's step label still says "Calling RAxML-EPA to obtain branch labels" even on this path, but the actual tiny.raxml.log timestamp is untouched from our own earlier RAxML-NG run — meaning EPA-ng ran, not RAxML, and the log message is just a stale leftover string (same "renamed the code but not the print statements" pattern as the stale version number). Cosmetic, but worth a one-line fix on his end too since it actively misleads about which engine ran."

Ok just pushed the new version with some fix : •

-version

flag added, and the banner now reports the fork version and the SATIVA base it comes from. epac/version.py had never been bumped since the fork. •

-reftree

without

-refmodel

is now an error instead of a silent GTR+G. Same bug class I'd fixed for

-r

and missed here. • The trainer's step labels name the engine that actually ran

Great. There's a queue for bioconda of course. Maybe it's best if I let Claude base the first steps on your repo rather than bioconda so we can find as many quirks as possible before you submit another recipe PR?

Sounds good, I've closed the PR. Just tell me when you're done finding quirks and I'll do a new one