Can someone look over my first schematic before I order boards?

A backwards catch diode doesn't wait for the coil to kick. The diode only does its job when the transistor shuts off. If it's reversed, the problem shows up the moment you energize the relay: the diode conducts forward from the 12v rail through your transistor to ground, so it's basially a short across the supply while the coil is on. That ussually kills the diode first, then the transistor right behind it, and sometimes takes the trace on the board with it. The good news is you can catch all six before power-up with a multimeter in diode mode. The band end shold face the positive coil terminal on every one, since all your relays are presumably wired the same way, check one carefully and you can spot-check the rest in a minute. On grounding, at an amp or two of relay coils you're not in the territory where you need a formal star layout. What actually matters at your scale is keeping the coil return currents from flowing through the same copper your micro uses as its ground reference. If your single tie point is near the connector and the relay driver grounds come back to that same point rather than daisy-chaining through the micro's ground path, you're fine. The noise problems people warn about show up when you're switching tens of amps or doing fast PWM, not six coils clicking on and off.

For the buck section, follow the datasheet layout as closely as the example circuit, not just the schematic. Input cap right at the chip, short loop on the switching side, and decent bulk capacitance on the 5v output. Undersized output caps usually show up as the micro browning out when several relays drop at once, which is an annoying one to chase. Yes, post the netlist. Honestly for a first board, also post the schematic as a PDF or screenshot even if it's rough. Netlists catch connected-wrong errors but a visual pass catches things like missing pull-downs on the transistor gates or bases that a netlist looks perfectly happy about.

Agree with posting the schematic as a picture, not just the netlist. On my first board I had a transistor drawn with the pinout of a diffrent package, netlist checked out fine and I only caught it because someone zoomed in on the screenshot. Now I print the schematic and physically trace a couple of nets with a highlighter before ordering, sounds old school but it catches stuff the tools never will.

Reading this made me go double check my own little relay board. I had one relay click and the whole chip reset the first time I powered it, turned out the 5v wire to the chip ran right alongside the coil wires for a few inches. Moved it over and it never happened again, so physical spacing seemed to matter even before I understood why. Do the coil wires need to stay away from the chip's connections on the board layout too, or is that mostly a wiring thing outside the board?

Following this one since I'm in a similar boat, just with a little motor driver instead of relays. One thing that tripped me up on my first go was the transistors themselves, I picked parts that looked beefy but barely handled the coil current once I actualy checked the datasheet. Worth doing the math on what each coil draws before you order, I assumed and got lucky.

Also on the first power up I only connected one channel at a time, which saved me when one of my connections was backwards. Slower but way less painful than letting all six loose at once.

One thing that saved me on my first relay board was adding test points. A couple of pads on the 12v rail, the 5v rail, and one at each transistor collector. Cost me nothing and made debugging so much easier than poking at component legs with a meter tip. Also, when you first power it up, do it with the relays hanging in the air or at least with a current limited supply if you've got access to one. If a diode is backwards like the other poster said, you want to find out before it's soldered onto copper.