prompt-injection-firewall

De la autor

Expert guide for defending AI applications against prompt injection, jailbreak attacks, indirect injection via tool outputs, and data exfiltration through tool calls — input sanitization, output filtering, canary token detection, and layered defense architecture / Panduan ahli pertahanan aplikasi AI terhadap prompt injection, serangan jailbreak, injeksi tidak langsung via output tool, dan eksfiltrasi data melalui tool calls — sanitasi input, filtering output, deteksi canary token, dan arsitektur pertahanan berlapis.