Realistically, how well will someone with a nonspecialist knowledge of biology be able to contribute to these projects?

That statement's not intended to pass judgement; I'd really like to know as it'd be great to contribute to a good cause.

Having helped pathology folks before (with absolutely no pathology knowledge), I would say testing and refactoring. Most of the people involved on these projects are not CS folk, so they write code that "works". Tons of repetitive, poorly tested, and usually inefficient approaches. Obviously you can't help much with domain specifics but a clean code base goes a long way. Or just fixing algorithmic or non-optimal approaches. For instance, fixing roundabout python code to use list comprehension, which is more idiomatic and more performant than most other raw python approaches.

A lot of bioinformatics specialists aren't talented programmers, they are biologists who know how to program as a means to an end.

As a computer scientist you can add a lot of value by contributing to optimization and ease of use.

Not to mention correctness, which is often sorely lacking.

Realistically, contributions from naive computer scientist (naive = no biology training) would be expected to be net negative in the medium term, as the cost of training them to be useful would come at the cost of daily productivity (same argument as Brook's Mythical Man-Month).

I've worked directly in the area covered by computer science and biology through most of my career. One anti-pattern I see that comes up over and over is an expert in CS will come along, see some article in Scientific American or wherever, and build a tool around some clever algorithm that doesn't really solve problems that are useful to scientists. Often the problem is lack of Deep Knowledge of Biology. I've found that the CS people (and physicists) who spend at least a couple years reading up on modern biology are much more effective contributors.

Also have some such experience, and I don't doubt this. On the flip side, watching biologists try to write software is mildly horrifying.

I'm a biologist who writes software, and it is horrifying.

You can contribute some money to pay for the specialist's time: both BioJulia and BioPython (through its supporting foundation) have support / sponsorship links.

Most of bioinformatics is pattern matching text files and performing statistics. You don’t need to know anything about biology beyond a rough hs level to contribute to the tooling, even just polishing existing workflows. Most tools are not written by computer scientists so there is a lot of optimization to be done in most pipelines.

Some initial reading would be to start with the blast algorithim and the sam file specification.

It should also be noted that basically all the modeling groups working on this are, at the moment, doing it for free. You could get in touch with the universities in question to see about giving to the labs directly.