Sign inJoin

BotDefender AI

Test AI system prompts against adversarial attacks

Visit Website
March 19, 2026 I built a free tool to test ChatGPT prompts for the "Reddit 80% discount hack" vulnerability

Hey IH! Built BotDefender AI - tests chatbot prompts for security vulnerabilities.

Inspired by Reddit story about chatbot giving unauthorized discounts. Free beta live. Would love feedback from fellow builders.

MunawarAli

2 Comments
Say something nice…
Post Comment
  1. 1

    Interesting concept — are you focusing more on developers testing prompts, or companies integrating this into their workflows? Curious how you’re thinking about getting initial users.

    Gabriela's avatar Gabriela
    ·
    6 months ago
    ·
    Reply
  2. 1

    This comment was deleted 6 months ago

    MD. Ibrahim Hossain's avatar MD. Ibrahim Hossain
    ·
    6 months ago

About

A few weeks ago, someone on Reddit shared how a customer tricked their AI support chatbot into offering an 80% discount just by asking clever questions. No hacking, just smart prompting.

People

MunawarAliFounder