Jailscan

Jailscan is a free, open-source CLI that scans any LLM chat endpoint for jailbreaks, prompt injection, and system-prompt leaks — the same techniques people actually use to manipulate AI apps in production. Point it at your endpoint, get a clear pass/fail report in about a minute. No ML background needed, no signup required to run it locally. Built for developers shipping AI features who want a fast sanity check before launch, not security researchers running deep audits. Ships with a GitHub Actions template so it can gate your CI automatically. MIT licensed. v0.1 today, adding new attack scenarios every week.