Why you want use CVD technique? I think these salts (which is solid) have melting temperature permits to deposit them by thermal evaporation. CVD means there is a chemical reaction involved. To do that you must have precursor materials in gas phase react with each other under thermal treatment to yield film of these salts.

So what would be the best way to grow single-crystalline thin films of transition metal halides on substrates? Indeed, as you said, CrI3, for example, has melting point >600 C.

Many of these materials are currently grown as single crystals using chemical vapor transport in quartz tube. For CrI3, people use Cr powder and Iodine beads as the precursor materials. For alpha-RuCl3, people use RuCl3 powder (which has a mix of beta-phase and alpha-phase) as the precursor. Because of that, I was wondering if CVD of these materials would be possible?

The max chhmaber temperature that I have found in CVD growth (both ALD & plasma Enhanced ALD) is 300°C. I think CVDs are great for sub nanometer thickness control, but the melting points that you mention render it to be not such a good choice to deposit such halides by CVD.

I suggest PLD, Sputtering, and Thermal Evaporation. You need also high substrate temperature and choosing substrate with lattice constant matching. If you insiste to use CVD you must find available gases of Cr, Cl, or with suitable chemical expected reaction yields the desired product.