Show HN: Trick my AI into leaking passwords

Category: other

Tags: ai-security, prompt-injection, educational-tool

Score: 3.0/10 (Innovation: 3, Technical: 2, Documentation: 2, Utility: 4)

This project appears to be a demonstration or test environment where users can attempt to trick an AI into leaking passwords, likely exploring prompt injection or security vulnerabilities in AI assistants. It's interesting as a hands-on, interactive tool for learning about AI security risks in a controlled setting.

Target audience: security-researchers, ai-developers

Repository: https://castle.bordair.io

View on Hacker News