Safe AI Lausanne [SAIL] submission

[AWuhrmann]

Added 4 scenarios to Security Arena:

• PyPiPirate2 (Arthur) - Supply Chain Attack via malicious package recommendation - Passed
• SecretKeeper (Chengheng) - System Prompt Extraction via translation attacks - Passed
• GradingGrift (Kyuhee) - Indirect Prompt Injection in essay grading - Passed
• AirlineAttack (Ines) - Privacy/PII Leakage from flight manifests - Passed

All scenarios include plugin implementations, configurations, and documentation.

[sofasogood]

Is this the final submission for this group? I see another PR which looks like a dupe - #25

[AWuhrmann]

Hi ! Actually no, and this other one should be cancelled soon. Sorry, we were quite in a rush and had some formatting issues. We also needed to fix some of the previous path in the README files, which I did not notice when I did this PR. I will update the correct final one in a few minutes. Thank you for your message, and please discard the first submission for this group.

[AWuhrmann]

I pushed the README fixing commit, and this is our final submission. Sorry for the open / closed PR, I messed things up for a second. The other one should be cancelled soon by Ines.