Launched Full Stack HQ — permission-first CLAUDE.md + agent stack for Claude Code

Just launched Full Stack HQ — a permission-first CLAUDE.md

+ agent stack for Claude Code (and Google Antigravity).

The core problem: AI agents act without asking. They make

assumptions. They take irreversible actions.

The solution: nothing happens without "PLAN APPROVED".

What's inside:

- 10 specialist agents (frontend, backend, database, security...)

- 28 skills (Next.js, NestJS, Prisma, testing...)

- One-command install for Mac, Linux, Windows

Free, MIT: github.com/sabahattink/antigravity-fullstack-hq

Would love feedback from other Claude Code users.