The calculations of the life support are completely busted, in addition to oxygen, in fact, food, water and the removal of CO2 must be added
https://ntrs.nasa.gov/search.jsp?R=20190027610

For Harry W. Jones of NASA, the life support material requirements 12.29 kilograms per crewmember-day:

Oxygen 0.84
Food solids 0.62
Food water content 1.15
Total water supply 9.68

At least these materials must be added to these resources to remove CO2

PS
I am a great fun of SpaceX, for me option A could work, but with the use of the Gateway that gives the following advantages:

- Reduces the deltav needed to enter the lunar orbit and to return to earth;
- Provides life support, the initial module (HALO) should provide 30 days of life support in addition to extra m3. To then increase with the I-HAB module (both modules built in part by Thales Alenia Space Italia);
- Allows the supply of goods (DragonXL) and propellants (ESPRIT module)

https://ntrs.nasa.gov/search.jsp?R=20190032114

Before the idea of launching the first two modules of the Gateway, the manufacturer of the PPE envisaged "Demonstrate xenon transfer capability"
https://www.thalesgroup.com/en/worldwide-space/or...

"ESPRIT, a propellant storage and refuelling system for the Gateway’s propulsion, as well as communications systems with the moon, interface points for external payloads and a scientific airlock."

I agree 100%

I would like to propose an objection: if you want to replace the SLS rocket with the Falcon Heavy and the New Glenn (Starship is a very promising joker, but with undecided times) you have to deal with the limitations deriving from the paylaod reduced to about 15 t in TLI. For this simple reason the gateway is not only necessary as a refueling and assembly point but it would take a second to act as a space tug between the Earth's orbit and the moon's orbit. Using an Isp of 2600 s (the AEPS engine) drastically reduces the propellant consumed and increases the payload delivered in lunar orbit. Delivery times increase to a few months but the advantages in terms of payload delivered are enormous;

Furthermore, if you want to exploit lunar resources in a sustainable way and, as the author says, to deliver thousands of tons, you need specialized logistics that makes the best use of existing technologies. Especially for goods this is possible by using the Gateway, I think so:
* launchers who deliver in Earth orbit (in this role starship is perfect)
* tugs transfer payloads to lunar orbit
* hydrolox lunar landers make the route with the Gateway

The tender for the Gateway Module (PPE) gives NASA the possibility to buy more modules.

Why not use solar electric propulsion (SEP) for LEO to NHRO transfers?
https://ntrs.nasa.gov/search.jsp?R=20180006894 https://ntrs.nasa.gov/search.jsp?R=20190000484
In these links the NASA analyzes the extra capacities of SLS 1B and those of the Falcon 9 and Atlas V if the SEP is used.

On page 15 and 16 it says:

"CONCLUSION
Two options for low thrust orbit insertion into an NRHO have been presented. The first option launches as co-manifested payload on SLS, while the second option is to launch on a CLV to an EEO and use a SEP spiral to the NRHO.

For the first option, launching as a co-manifested SLS payload, a spacecraft with a 6,500 kg wet mass can be inserted into NRHO with as little as 90 kg of propellant in 76 days.
[...]
However, the trades between mission constraints such as NRHO mass, TOF, V, and BOL operating power must be considered when choosing the most appropriate insertion trajectory and CLV. It was found that the AV can enable a maximum NRHO mass of 9,500 kg in 500 days TOF, and the F9 can enable a maximum NRHO mass of 6,500 kg in approximately 380 days TOF.
If a total TOF of less than 1-year is desired, the maximum NRHO mass is 8,100 kg and 6,200 kg for the AV and F9, respectively. If only 1,000 kg of propellant can be used, NRHO masses of approximately 7,200 kg (AV) and 5,500 kg (F9) are possible."