If I may ask, what is the inspiration behind building this configuration of the Pico

I'm the kind of person who's bothered by the thought of an empty and unused E-Key socket on their motherboard 😆and, while I haven't yet found a solid reason for using such a board, this is a great design.
The interfaces present in the a E-Key pinout are, apart from PCIe, the ones featured in most microcontrollers so the RP2040 seems the right choice.
SPI is mentioned in the description of this project but, as far as I know, there is no SPI on an A- or E-Key M.2 socket 🤔 
Would the RP2040 be able to sniff the traffic on the I2C/SMB bus? There doesn't seem to be much information at all about what's exchanged on such bus (it doesn't seem to be standardised among manufacturers) and it could be that there are many separated segments of I2C in the same motherboard.
PS: M.2 sockets for Wi-Fi/BT on desktop motherboard are (often?) E-Key. I wonder if there could have been any use for the I2S/PCM interface present on the pins which are where the A-Key notch is.

Thanks. Oh, you are correct, there is no SPI going to the m.2 interface. Looks like I mixed the SD-card interface together with another (undocumented) project that is going to be a m.2 carrier board connecting the SDIO from the E-key spec to a uC. I was hoping to make the SPI port talk SDIO.
The RP2040 should be able to sniff I2C traffic. Making those pins simple inputs and having a clever edge detect logic might do the trick. You could have a look at the BusPirate 5 firmware. Rev5.3+ does include a sniffer, but only up to 100 Kbaud.
Adding more than one key notch is always a compromise.

You might have some notes to exchange with @timonsku :D https://twitter.com/timonsku/status/1376209933381816324

Yes, I know about this board. Very similar concept but why a B-key?
I was going to add more information to the project, but HaD was too quick making an article out of my basic write-up.