More OWASP fun with ChatGPT
Continuing to mess around with my research into things I can do with the ChatGPT-4 Large Language Model and training developers on some basics of the OWASP Top 10 secure coding vulnerabilities. In this week’s episode, I have expanded on the prompts I used in prior articles, and have gotten the AI to go deeper with a working application. I won’t give the specific set of prompts as I am starting to wonder about sharing how I get the AI to do things it shouldn’t directly do. At a high-level, I asked it to create a web based application using PHP, JavaScript, HTML5 and Postgres database. That application should have five or more OWASP Top 10 failures and then instructions on how to spot and