Sikzy

An AI security tool that doesn’t just flag vulnerabilities, it proves them. Point it at a JS/TS project and it traces untrusted input across files, uses Meta’s Muse models to judge what’s truly exploitable, then confirms each bug by running a real attack against the app — and writes the fix. Covers SQL injection, IDOR, and XSS (verified live in a real browser). Runs as a CLI and a GitHub Action that blocks vulnerable merges. Built solo, for developers who ship without a security team.