You can opt for hypergolic propellants that won't boil off, but you sacrifice performance and make ground handling more expensive and dangerous due to these being highly toxic.
And now my mission profile has two different landing modules, one using cryogenics, one using hypergolics. Oh well... We can always go hypergolic if the mass budget is not a big issue.
Apollo basically had a switch + sensor based tigers to start an automatic segvence that shuts down and drops the first stage, starts the seconds stage and flies it back to lunar orbit.
If you don't have an ascent stage or enough delta-v, you have to land, potentially many kilometers of course or even crash if something goes very wrong.
That's a good reason to have multiple supply/equipment depots on the surface before you even land the ascent vehicles. Walk up to the nearest one, then drive the rest of the way.
Also that's one good reason to have spare ascent craft (or supply depots) parked on low lunar orbit that could be directed to pick up any stranded astronauts.
A last option is a LESS-like lifeboat assembled from lander components. Surface to orbit on a convertible has some appeal.
Already considered in Apollo days:
https://en.wikipedia.org/wiki/Lunar_escape_systems
A very simple "rocket chair" using residual fuel from the LEM. There were also variants for regular use during exploration using only leftover propellant from the descent stage.
Flown manually all the way to lunar orbit due to guidance computer being to heavy back then:
"Guidance in typical LESS designs was simple: an 'eight-ball' to show spacecraft attitude, a clock to show time since liftoff, and a planned pitch program. The Apollo Guidance Computer used as an autopilot for the CSM and LM had a mass of around a hundred pounds and consumed a significant amount of power, so computer controlled flight was out of the question. This would be one of the few cases where an astronaut flew a rocket manually all the way to orbit, and with far less instrumentation than normal."
It can be done, but can it be guaranteed?
True. No wind or waves, everything standing perfectly still.
I'd say it's much easier.
Sure, we have much better radar and composition maps of the lunar surface these days, but it strikes me that it's still not as 100% as landing on a barge's [presumably] metal surface or concrete and Cape Canaveral.
I'm more worried about sandblasting the surroundings.
I'm sure smarter people than I have thought about this, but it was a thought that struck me when I saw the mockups in the NASA lunar bid vid a wee while back.
Still, by then the ship should be somewhat top-heavy and those six landing pinkies do not inspire much trust.
OTOH, the people designing this are smarter than we are and much more experienced than me in rocket design, so they may know one or two things I don't.
And American astronaut is launched in a hurry due to the space race in a one-way Gemini based lander, lands off course and needs to find the shelter/ascent vehicle with limited supply of air. And at the same time the Soviet mission he was effectively racing with goes silent...
You can launch the lander first, park it in lunar orbit, then launch the crew on a second Falcon Heavy, each having 38% of an Apollo worth of mass. With a little more creativity, you can launch the ascent vehicle on another Falcon Heavy and land it before the other two so that astronauts can rest assured that, unless they kill themselves during the landing, there is a ride back up waiting for them.
With that you can have 3 separate chronograms and each mission is simplified in relation to a single Apollo.
If you are unhappy with the mass budget, you can launch yet another series of missions with supplies and equipment to be landed before the ascent vehicles are landed. And, since in this case whatever lands doesn't even need to be pressurized, or delicate, landing it should be much easier too.
Now we are at 1.52 Apollo masses per mission. It took 4 launches to get there but, still, it's a lot of stuff for our astronauts to work with. Also, if any one launch fails, you can always try again.