(I'm reading between the lines to assume that you don't actually have a problem with teams standardizing around a specific stack, but rather that they shouldn't market themselves that way; either way, that's not really the interesting part.)
Your counterargument is going to be "don't do government stuff" I'm sure.
That's a rule that has served us in pretty good stead.
It's not just the moral issue (though I do have dealbreaking moral problems with doing the kind of work I do for the USG) but also that the market for GSA-style services is so different than the regular market; not only is it harder and weirder to address as a market, but also succeeding there doesn't teach you much about how to make your offerings work in the real market.
(I've worked, earlier in my career, on somewhat large-scale projects [not in vuln research] serving government agencies, but have never been a principal at a company that targeted government sales, so, grain of salt or whatever).
GSA products sales aren't so bad, and sometimes services are incidental to product sales. If I were a lock maker, I'd sure want to be Kaba-Mas or S&G with a decent amount of government sales. If you have to provide some services, it's not horrible -- it's a $200-400k/yr living for a few consultants if you're in the right space, and generates nice revenue for the company, without taking over the focus. If you're an independent contractor and end up subcontracting to a product company, you can make really good money filling this role. If you don't have to do the "get the product sold" phase, and are purely subcontracted to the the implementation/support, it's actually not bad, and not very different from enterprise clients in general. There is the problem of single-client-all-revenue, and some slightly odd work rules, but enterprise often has that.
I share your moral concerns about vuln research/etc. for the government, but I have no problem with defensive tech for them, or for non-security tech.
An hour or two of ancillary work? Do it as a favor, not a billable project.
Negotiate scope, not rate. Try to make rate immutable.
You are charging too little.
Your rate is probably not a straightforward function of what your fulltime equivalent makes.
Get an accountant now.
Have a well-defined master contract and a separate statement of work (SOW) for each project.
Get acceptance criteria into the SOW, but don't make it ultra-specific or you'll invite arguments.
Incorporate. It's cheap insurance.
Get a lawyer and have your contracts reviewed.
Expect to work on your customers' paper, not yours. See: get a lawyer.
Reach out periodically to all your not-unsatisfied clients. You'll be surprised how much work this drums up.
Specialize on domain, not on technology stacks.