AI agents aren't safe from prompt injection, and spreadsheets prove it

3 points | by taubek 8 hours ago

1 comments