We in fact have a standard method to test whether sequence is under selection or not, and engineering is a kind of selection. It is called Ka/Ks ratio.

https://en.wikipedia.org/wiki/Ka/Ks_ratio

For SARS-CoV-2, there were 79 amino acid changes and 475 silent changes relative to known common ancestor. This is a strong signature of purifying selection, hence against any engineering. The article does not address this evidence at all.

You can argue that sequence was engineered specifically so that it doesn't look engineered. That's impossible to refute with sequence evidence, but also not compatible with accidental release.