GitHub - sabahattink/antigravity-fullstack-hq: Production...

Full Stack HQ

The most opinionated AI coding configuration for serious engineers.

Works with Google Antigravity IDE and Claude Code — out of the box.


`GEMINI.md` + full agent stack	`CLAUDE.md` + full agent stack

The Problem

AI coding agents are powerful — but without guardrails, they become unpredictable. They create files without asking. They make assumptions. They break things mid-session.

Full Stack HQ solves this.

Install once. Your agent becomes a disciplined senior engineer who:

✅ Always plans before acting
✅ Uses the right specialist for each task
✅ Knows your full tech stack deeply
✅ Never surprises you with unexpected changes
✅ Writes consistent, production-grade code every time

What's Inside

GEMINI.md — Global rules for Antigravity

Permission-first workflow
Tech stack defaults
Code style enforcement
Git conventions

CLAUDE.md — Global rules for Claude Code

Same philosophy, Claude-native syntax
Agent role definitions
Slash command workflows
Security checklist

Component	Count	What it does
Agents	10	Specialist AI personas for each domain
Skills	28	Deep knowledge modules (Next.js, NestJS, Prisma...)
Workflows	10	Slash commands for your dev loop

Install in 30 Seconds

Mac / Linux

curl -fsSL https://raw.githubusercontent.com/sabahattink/antigravity-fullstack-hq/main/install.sh | bash

# Or clone for more control
git clone https://github.com/sabahattink/antigravity-fullstack-hq.git
cd antigravity-fullstack-hq
./install.sh

Flag	Effect
`--only-antigravity`	Skip Claude Code
`--only-claude`	Skip Antigravity
`--force`	Overwrite existing configs

Windows (PowerShell)

irm https://raw.githubusercontent.com/sabahattink/antigravity-fullstack-hq/main/install.ps1 | iex

# Or with options
.\install.ps1 -OnlyClaude
.\install.ps1 -OnlyAntigravity
.\install.ps1 -Force

After install, restart your IDE. That's it.

The Permission-First Workflow

Every action requires your explicit approval. The agent plans, shows you what it intends to do, and waits.

You:    "Add user authentication with JWT"

Agent:  Here's my plan:
        Phase 1: Create auth module + JWT strategy
        Phase 2: Add guards to protected routes
        Phase 3: Implement refresh token rotation

        [APPROVAL NEEDED] Should I proceed with Phase 1?

You:    PLAN APPROVED

Agent:  [implements Phase 1 only, then stops and reports]

Valid approval keywords:

PLAN APPROVED  ·  IMPLEMENTATION APPROVED  ·  PROCEED  ·  DO IT

Anything else = the agent waits.

Tech Stack

Full Stack HQ is tuned for modern production stacks:

Frontend     Next.js 15 (App Router)  ·  TypeScript 5  ·  Tailwind CSS v4
Backend      NestJS  ·  Node.js 22+  ·  BullMQ  ·  Redis
Database     PostgreSQL 16+  ·  Prisma 6+
Auth         JWT with refresh token rotation
Testing      Vitest  ·  Jest  ·  Playwright
Infra        Docker  ·  GitHub Actions  ·  Vercel

Not using this stack? Edit ~/.claude/CLAUDE.md or ~/.gemini/GEMINI.md to swap any technology.

Agents

View all 10 agents

Agent	Trigger phrase	Specialty
`frontend-specialist`	`Use the frontend-specialist to...`	React, Next.js, Tailwind, UI/UX
`backend-specialist`	`Use the backend-specialist to...`	NestJS, APIs, queues, Redis
`database-specialist`	`Use the database-specialist to...`	Prisma, PostgreSQL, migrations
`architect`	`Use the architect to...`	System design, ADRs, trade-offs
`code-reviewer`	`Use the code-reviewer to...`	Quality, patterns, security
`test-engineer`	`Use the test-engineer to...`	Vitest, Jest, Playwright
`security-auditor`	`Use the security-auditor to...`	Auth, OWASP, input validation
`devops-engineer`	`Use the devops-engineer to...`	Docker, CI/CD, deployment
`performance-optimizer`	`Use the performance-optimizer to...`	Bundle, queries, rendering
`documentation-writer`	`Use the documentation-writer to...`	Technical writing, READMEs, ADRs

Workflows

Command	When to use
`/plan`	Before starting any feature — produces a phased breakdown
`/brainstorm`	Exploring architecture options before committing
`/debug`	Systematic root-cause analysis on bugs
`/create`	Implementing an approved plan
`/enhance`	Improving existing code quality
`/test`	Generating or fixing tests
`/status`	Progress checkpoint on long tasks
`/orchestrate`	Multi-agent task coordination

What Gets Installed

~/.gemini/
├── GEMINI.md                     ← global rules for Antigravity
└── antigravity/
    ├── agents/                   ← 10 specialist agents
    ├── skills/                   ← 28 skill modules
    └── workflows/                ← 10 workflow files

~/.claude/
├── CLAUDE.md                     ← global rules for Claude Code
├── agents/                       ← 10 specialist agents
└── skills/                       ← 28 skill modules

IDE Support

Feature	Antigravity	Claude Code
Global rules file	✅ `GEMINI.md`	✅ `CLAUDE.md`
Specialist agents	✅	✅
Skills / modules	✅	✅
Slash workflows	✅	✅
Hooks system	❌	✅
MCP servers	❌	✅
Plan mode	❌	✅

Contributing

Found a missing skill? Have a better agent prompt? PRs are welcome.

git clone https://github.com/sabahattink/antigravity-fullstack-hq.git
cd antigravity-fullstack-hq

# Add a new skill
mkdir skills/your-skill-name
echo "# Your Skill" > skills/your-skill-name/skill.md

# Open a PR
git checkout -b feat/add-your-skill

Please read CONTRIBUTING.md before opening a PR.

Credits

Built with inspiration from the community:

vudovn/antigravity-kit
sickn33/antigravity-awesome-skills
Vercel Labs official skills
Anthropic official skills

推荐订阅源

Hacker News: Show HN