Expand the number of injection strings

[puhley]

This is a catch-all for an issue that will have child bugs for each injection type. There is a balance that needs to be maintained in terms of the number of injection strings in the attack files:

• If you have too few injection strings, then things will be missed.
• If you have too many injection strings, then you will generate more images than can reasonably be tested by human teams.

Therefore, this project needs to create a curated list of highly successful attack strings that are enough to signal that there is was an issue with their application. At the same time, the list can't become so long that people avoid using this tool because it takes too long or it is too much work to test that many generated images.

The current lists are too short and we need to collect more highly successful strings.