AI coding agents can move quickly. The engineering challenge is keeping intent,
approval, and verification visible while they do.
That is the problem Full Stack HQ is designed to address: a shared,
permission-first engineering workflow that can travel across multiple AI
coding hosts.
The project supports Claude Code, Google Antigravity IDE, and OpenAI Codex from
one tool-agnostic source of truth.
Most people configure their AI agent once (or never) and just... let it go. The result is an agent that:
The agent is powerful but unpredictable. That's the worst combination in software development.
The installed rules are designed to make the agent plan, show what it intends to do, and request explicit approval before execution.
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]
The only valid approval keywords:
PLAN APPROVED
IMPLEMENTATION APPROVED
PROCEED
DO IT
The rules define these as the approval phrases. They are prompt/configuration guidance, not a runtime permission boundary, so the host agent still determines the observed behavior.
| Component | Count | Description |
|---|---|---|
| Shared rule core | 1 | Host-neutral engineering policy |
| Host adapters | 3 | Claude Code, Antigravity, and Codex |
CLAUDE.md /GEMINI.md /AGENTS.md |
3 | Generated host instruction files |
| Agents | 10 | Specialist AI personas |
| Skills | 28 | Domain-specific knowledge modules |
| Workflows | 10 | Legacy workflows plus skill bridges |
Instead of one generic agent trying to do everything, you get domain experts:
| Agent | What it handles |
|---|---|
frontend-specialist |
React, Next.js, Tailwind |
backend-specialist |
NestJS, Node.js, APIs |
database-specialist |
Prisma, PostgreSQL, migrations |
architect |
System design, trade-offs, ADRs |
code-reviewer |
Quality, patterns, best practices |
test-engineer |
Vitest, Jest, Playwright |
security-auditor |
Auth, OWASP, input validation |
performance-optimizer |
Bundle, queries, rendering |
devops-engineer |
Docker, CI/CD |
documentation-writer |
READMEs, technical writing |
Calling them is simple:
Use the database-specialist to design a user schema with soft deletes.
Deep knowledge modules for the tools you actually use:
nextjs-app-router, react-best-practices, ui-ux-pro-max, frontend-design
nestjs-patterns, prisma-workflow, software-architecture
test-driven-development, systematic-debugging, webapp-testing
brainstorming, prompt-engineering, skill-creator
/plan → phased breakdown with approval checkpoints
/brainstorm → explore architecture options
/debug → systematic root-cause analysis
/create → implement an approved plan
/enhance → improve existing code quality
/test → generate or fix tests
/orchestrate → coordinate multiple agents
/ui-ux-pro-max → structured UI/UX review
The canonical workflow bodies remain in `workflows/`. The installer keeps the
Antigravity legacy form and also renders each one as a skill for Claude Code,
Codex, and modern Antigravity.
The installers copy files from the cloned repository and should be run from that checkout.
Mac/Linux:
git clone https://github.com/sabahattink/antigravity-fullstack-hq.git
cd antigravity-fullstack-hq
chmod +x install.sh
./install.sh
Windows (PowerShell):
git clone https://github.com/sabahattink/antigravity-fullstack-hq.git
Set-Location antigravity-fullstack-hq
.\install.ps1
Options:
./install.sh --only-claude # Claude Code only
./install.sh --only-antigravity # Antigravity only
./install.sh --only-codex # OpenAI Codex only
./install.sh --dry-run # Preview without writing targets
./install.sh --force --backup # Replace managed files with backups
The scripts perform pre-flight checks and install the selected host files.
~/.claude/
├── CLAUDE.md ← global rules (Claude Code)
├── agents/ ← 10 specialist agents
└── skills/ ← 28 skill modules
~/.gemini/
├── GEMINI.md ← global rules (Antigravity)
└── config/
├── agents/
├── skills/
└── workflows/ ← legacy bridge
~/.codex/
├── AGENTS.md ← global rules (Codex)
└── agents/ ← TOML custom agents
~/.agents/skills/ ← shared Codex skills
The shared rules describe several things I found critical in practice:
1. Separation of planning and execution
The rules separate planning from execution: the agent should plan, request approval, then execute the approved slice.
2. Role-based reasoning
Before acting, the agent asks: "Who is the right specialist for this?" A database schema question goes to the database specialist, not the frontend agent pretending to know Prisma.
3. Explicit code style
No semicolons. Single quotes. 2-space indentation. Arrow functions. Named exports. These are documented defaults that the host agent is asked to follow.
4. Security checklist
Before every commit, the rules include a checklist for hardcoded secrets, input validation, bounded queries, and rate limiting. The checklist is guidance for the host agent, not an automatic scanner.
The mental model I was missing: AI agents should behave like senior engineers, not interns with root access.
Senior engineers don't start typing when you describe a problem. They think, propose a plan, get sign-off, then execute — one reversible step at a time.
Full Stack HQ encodes this discipline in its global rules and workflow files.
⭐ github.com/sabahattink/antigravity-fullstack-hq
MIT license. Open to PRs — especially new agents and skills.
What does your current shared agent configuration look like? I'd love to see
what rules others have found valuable across their preferred hosts.