Free defensive tools

Assess the LLM app, not just the prompt.

Each Rogue Prompt tool runs entirely in your browser. No signup, no backend, and no pasted prompt or architecture note leaves your device.

LLM security checklist

Prompt Injection Risk Checklist

Score an LLM app by untrusted inputs, tool authority, data exposure, and human review controls.

Open tool

System prompt leak prevention

System Prompt Hardening Linter

Paste a system prompt and get client-side defensive findings about secrets, brittle wording, and missing boundaries.

Open tool

AI security assessment template

LLM App Threat Model Generator

Select an app shape and generate a markdown threat model covering surfaces, mitigations, and test cases.

Open tool