Troiana Prompt Library · Claude · Development

Red-team a security-sensitive change

Make Claude attack the change before someone else does.

What this prompt helps you do

A risk-ranked list of attack paths with preconditions, impact, and concrete mitigations.

When to use it

A change touches authentication, permissions, payments, or untrusted input.

What to prepare

The goal, relevant source material, constraints, examples, and the decision or action this should support.

How to get a strong result

Ask for attack paths with their preconditions, not a checklist of general security principles.

How to use this prompt

Replace each placeholder with accurate context. Paste the complete prompt into Claude, review the response critically, resolve important questions, and use it as a considered input to the next decision—not as an unchecked final answer.

Browse all ChatGPT, Claude, and Codex prompts