BestPromptFinder

Threat Model a New Feature Before You Build It

Quality, usefulness and confidence are AI evaluations out of 100 — not user ratings. Community results come from visitors' "worked / didn't work" votes below. How scoring works →

The prompt

Run a lightweight threat model on a feature I'm about to build. Feature: [e.g. users can upload a profile photo / invite teammates by email / pay by UPI / export their data] How it works: [the flow step by step, which services and databases it touches] Users and roles: [who can do what] Data involved: [personal data, payments, files, tokens] Stack: [ ] Work through: 1. A simple data-flow description: where data enters, where it's stored, where it leaves, and the trust boundaries. 2. Threats by category (spoofing, tampering, information disclosure, denial of service, elevation of privilege, and abuse of the feature itself), each with a concrete example for this feature. 3. A table: Threat | Likelihood (low/med/high) | Impact | Mitigation | Effort. 4. The five mitigations to build in from day one, and test cases that prove each works. Rules: be specific to my flow, not a generic checklist; mark assumptions clearly; don't suggest anything that would break the feature's purpose without saying so.

Why this prompt

• Hand-picked, editorially reviewed prompt
• Written for Claude, ChatGPT, Gemini

Source

Created by BestPromptFinder — editorial prompt, not user-generated. See our content & source policy.

Related Coding prompts