[feature] ability to reverse moon ci task logic

[shellscape]

Is your feature request related to a problem? Please describe.

By default moon ci runs everything and it's up to us to annotate everything we don't want it to run. We've got a lot of tasks we don't want it running, and conversely only have a few that we want it to run.

Describe the solution you'd like

I'd like the ability to opt-in to moon ci running a task, rather than opt-out, as the opt-out is burdensome in a few repos I'm working in.

Describe alternatives you've considered

Aren't really any other ways to get around this one at present.

Additional context

In the repos I'm using moon, fine grained CI control and setup is preferred over letting moon ci run all tasks. Linting, building, etc are preferred in separate steps with separate reporting in Github Actions. The DX for that is viewed as better than having a single task step output the results for steps which are seen as separate, incremental, blocking steps. e.g. We perform linting first. Linting blocks Building. Building blocks Deployment. A failure in the previous step prevents the following steps from executing. While it might not be everyone's cup of tea, it's what we like and serves us well. A single step with output for all-the-things is hard to read, and hard to parse in a repo wit 20+ projects. We also lint the entire repo and cache eslintcache in the workflow to improve speed. We do this because it's a safer bet.

Our end goal is to have moon ci run only our deployment step, as that's the only part of our task system that requires knowing what the affected files are.

[milesj]

Not opposed to it, will look into 👍

[milesj]

@shellscape This should do it #1056

[milesj]

Can pass targets now in v1.14!

[shellscape]

Huge!