Implement shellcheck for code quality

[jorgepiloto]

ShellCheck is a great tool for ensuring consistency and best practices across shell scripts. It expects complete shell scripts but I bet we can create a custom workflow for checking shell: bash blocks.

It will save us from issues like this one: #901

[moe-ad]

While this is a great idea in principle, the time I spent investigating how to proceed revealed that a robust + practical implementation will be very complicated and hacky. By practical, I mean an implementation that will actually help catch the cases that give us the most headaches in ansys/actions. ansys/actions steps are very dynamic in nature, meaning there is a lot of runtime evaluation of data e.g. interpolation happens for ${{ ... }} contexts, python scripts set environment vars, steps communicate with each other through env vars etc. Most of the issues/headaches we face stems from this core nature of ansys/actions.

Running shellcheck against parsed "run" steps (as is currently being explored in #906) will result in a lot of false positives with respect to all these dynamic data (since shellcheck is not designed for actions and will not be able to correctly handle github syntax related to dynamism), unless we add some workarounds that basically boils down to skipping areas of our actions related to dynamic data (we end up with something similar to what https://github.com/saleor/shellcheck-gha does), which defeats the whole purpose of what we want to achieve.

But maybe that is sufficient as a starting point, and we can iterate in the future. But I would rather urge us to think about even more ways we can improve our end-end testing, similar to #929 for example.

Eitherways, I am removing this from v10.3 milestone and adding it to v11 so we can have more time to think about this.

Pinging @SMoraisAnsys, @MaxJPRey, @RobPasMue, and @jorgepiloto for visibility.