By moving the linter to its own job, we don't need to block test runs on minor linter errors. This lets contributors first get their work done without needing to polish and fix linter errors constantly throughout.

Running the linter in parallel also saves 30s to 3 minutes (linter time varies).

I've kept the linter requirement for dev-release as it should mirror prod-release as much as possible. And of course the linter (and every other job) should pass before we run a prod-release.

So that we can see all linter errors at once and fix them together.

Changing npm-run-all to serial adds 10 seconds, but prevents linters from competing for resources. It also provides better progress output without needing to aggregate.

chore(ci): run lint and build in parallel

A previous change moved the filtering out of base image vulns (based on
the existance of `dockerfileInstruction`) to be before
`dockerfileInstruction` is assigned to the vulns, so all vulns were
filtered out in case we used the `exclude-base-image-vulns` option.

This fix moves the filtering back to where it was before, and
recalculates the value of `ok`, to account for the reason it was moved
in the first place.

docs: update log4shell command help

docs: update monitor command help

docs: update config command help

docs: update container command help

docs: update iac command help

Updates to test command help

fix: exclude base image vulns filtering all vulns