Gemini Adversarial Prompt Analyzer
8.2
A tool for security researchers and AI developers to analyze and categorize adversarial prompts targeting large language models like Gemini. The software will collect, tag, and analyze prompts to identify patterns and vulnerabilities used in cloning attempts and other malicious activities, allowing for proactive model hardening.
240h
mvp estimate
8.2
viability grade
5
views
technology stack
Python
SQLite
Difficult
inspired by
Attackers prompted Gemini over 100,000 times trying to clone it