I tried to make the most consistent, vanilla version of the deck in the list above using cards I currently have available (you could definitely run 4x underrealm lich). I want to keep lotleth giant since he's the only source of direct damage in the deck, I ended up dropping carnage tyrant for the 4th midnight reaper. I want to run a full complement of creatures to maximise undergrowth payoffs. Maximising the count of creature removal creatures will help out your plaguecrafters a bunch by wittling down the enemy ranks, perhaps stripping protection from a planeswalker. Guildmage forum is optional, I like it because it can pump a glowspore to 4/2 and out of chainwhirler range. Glowspore can even reach 5/3 for 4 mana with two forums - in mtg: arena make sure you tap both mana you want to convert first, then use the forums, otherwise you can waste your special, converted mana, indicated with white diamond on the mana icon (the forums use the oldest mana in the mana pool first by default).

The way I would use the maybeboard to make a deck variant would be to first choose what I'm cutting from the deck to make room. First of all I wouldn't touch the core self-mill engine of 4x stitcher's, 4x glowspore and 2x underrealm lich. If you cut the other 4 of cards down to 3, that makes space for 6 new cards. You could go further and then cut all the 3 ofs down to 2, making room for 12 new cards in total. I've also tried making a 90 card deck, XL Rainbow Hatchery Spider - this allows you to add 18 cards and 12 land. More cards means less consistency, but perhaps the wider selection of utilty creatures and the larger potential undergrowth payoffs could make it worthwhile - at the very least it's a good way to experiment with a wider variety of cards.

Here's the first run I tried in standard, going 2-3. The variant I played ran a utility creature toolbox including 2x nullhide ferox, 1x pilfering imp, 1x thrashing brontodon, 1x plague mare. I was only running 1 ravenous chupacabra and 1 kraul forager, which now makes me wince a little. Perhaps the utility approach in best of one is inherently weak, as you often draw the wrong tool, the next attempt I might try the more generic, vanilla deck above.

Dec 13th edit

I tested the vanilla undergrowth deck in constructed - it went 1-3 then 7-1. First run I met a couple of angel decks running lyra dawnbringer, a problem because the first strike means she's safe from kraul harpooners. Second run felt great, but some opponents definitely had weak draws, the monoblue deck I played didn't get curious obsession or tempest efreet the entire game and seemed short of counterspells.

Dec 17th edit

I had a 3 hour ranked session with the deck. I should probably add timestamps for the start of each game.

Further reading and other vids:

The Firemind's GP Milwaukee PTQ top 8 report - reddit thread with a lot of interesting discussion.

Gabriel Nassif's in depth article on undergrowth - shows how his deck list evolved through testing, has notes and sideboard suggestions for different matchups.

JackischMTG's undergrowth deck that went 17-4 in event queues

Hinterland Sage - he's running 20 lands with explore creatures, pelt collector's, llanowar elves, golgari raiders, necrotic wound and assassin's trophy. It's a lower to the ground, aggro approach.

Jim Davis played undergrowth in the kibler's giant monsters event, then tried it in standard.