[Feature request]: thinking we should get over the prompt injection content from Pliny the Elder.

[danielmiessler]

Repos like this, for example.

https://github.com/elder-plinius/L1B3RT4S

And I want to be very careful about giving credit here in the main Readme and also in the directory where we put the files.

If we know how to reach out to them, we should also ask for permission beforehand. But I'm pretty sure the answer is yes, because it's a public repo. It's only going to give them more exposure to have it here as well. But I want there to be very clear pointers that point to their work and that give them credit.

[ItsIgnacioPortal]

I've been looking at the L1B3RT4S repo, and it seems to have many wordlists with no clear objective, design principles, nor any documentation. For AI jailbreaks wordlists we should consider other projects, like: https://github.com/General-Analysis/GA

It would be no problem to incorporate new wordlists into the project, as each project shared within SecLists is kept simply as a mirror, while keeping a direct link to the original repository in the documentation. We've done the same for the Trickest wordlists and many others. It should also be no problem to make it update automatically, since we already have working GitHub Actions workflows for that.

[ItsIgnacioPortal]

Also consider: https://github.com/verazuo/jailbreak_llms