Everything Antigravity

The definitive collection of configurations for the Google Antigravity ecosystem.

Unlock the full potential of your Antigravity assistant with production-ready agents, skills, hooks, commands, rules, and MCP configurations. Fully optimized for Antigravity 2.0's Dynamic Subagents, Asynchronous Automation, and Enterprise Integrations. Battle-tested patterns evolved from real-world development.

Supported Ecosystem

These configurations are 100% compatible across the entire Antigravity suite:

• Antigravity 2.0: The core agentic platform powering parallel subagents and automation.
• Antigravity IDE: The standalone desktop app for visual development and code management.
• Antigravity CLI: The fast, terminal-first interface for local orchestration.
• Antigravity SDK: The developer kit for building custom programmatic agents.

What's Inside

This repository provides a comprehensive suite of tools designed to enhance your workflow with Antigravity:

everything-antigravity/
├── agents/           # Specialized subagents for delegation
│   ├── planner.md           # Feature implementation planning
│   ├── architect.md         # System design decisions
│   ├── code-reviewer.md     # Quality and security review
│   ├── tdd-guide.md         # Test-driven development
│   ├── security-reviewer.md # Vulnerability analysis
│   ├── build-error-resolver.md # TypeScript/build error fixing
│   ├── e2e-runner.md        # Playwright E2E testing
│   ├── refactor-cleaner.md  # Dead code cleanup
│   └── doc-updater.md       # Documentation sync
│
├── skills/           # Workflow definitions and domain knowledge
│   └── coding-standards/    # Language best practices
│
├── commands/         # Slash commands for quick execution
│   ├── tdd.md               # Test-driven development workflow
│   ├── plan.md              # Implementation planning
│   ├── code-review.md       # Code review
│   ├── build-fix.md         # Incremental build error fixing
│   ├── checkpoint.md        # Workflow checkpoints
│   ├── e2e.md               # E2E test generation and execution
│   ├── eval.md              # Eval-driven development
│   ├── learn.md             # Extract reusable patterns
│   ├── orchestrate.md       # Multi-agent workflow orchestration
│   ├── refactor-clean.md    # Safe dead code removal
│   ├── test-coverage.md     # Coverage analysis and test generation
│   ├── update-codemaps.md   # Architecture documentation
│   ├── update-docs.md       # Documentation sync
│   └── verify.md            # Comprehensive verification
│
├── rules/            # Always-follow guidelines
│   ├── security.md          # Mandatory security checks
│   ├── coding-style.md      # Code style and conventions
│   ├── git-workflow.md      # Git workflow guidelines
│   ├── testing.md           # Testing requirements
│   ├── patterns.md          # Common patterns
│   ├── performance.md       # Performance guidelines
│   └── agents.md            # Agent orchestration guidelines
│
├── contexts/         # Contexts for specific modes
│   ├── dev.md               # Development mode
│   ├── research.md          # Research mode
│   └── review.md            # Code review mode
│
└── skills/           # Workflow definitions and domain knowledge
    ├── coding-standards.md  # Language best practices
    ├── backend-patterns/    # Backend architecture patterns
    ├── frontend-patterns/   # React/Next.js patterns
    ├── tdd-workflow/        # TDD workflow patterns
    ├── security-review/     # Security checklist and patterns
    └── verification-loop/   # Verification system

Getting Started

1. Installation

Clone this repository into your project or integrate specific components into your Antigravity configuration.

git clone https://github.com/iamaanahmad/everything-antigravity.git
cd everything-antigravity

2. Configure Agents

Copy desired agents from agents/ to your .agents/agents/ directory (for project scope) or ~/.gemini/config/agents/ (for global scope).

# Copy all agents
cp agents/*.md .agents/agents/

# Or copy specific agents
cp agents/planner.md .agents/agents/
cp agents/code-reviewer.md .agents/agents/

3. Set Up Rules

Establish project-wide rules by copying AGENTS.md into your .agents/ customization root.

cp AGENTS.md .agents/AGENTS.md

4. Add Skills

Copy skills to enhance Antigravity's domain knowledge:

cp -r skills/* .agents/skills/

5. Configure MCP Servers

Copy MCP server configurations and replace placeholders with your actual API keys:

cp mcp-configs/mcp-servers.json .agents/mcp-servers.json
# Edit the file and replace YOUR_*_HERE with actual values

Key Concepts

Agents

Delegate specific tasks to specialized subagents. Take advantage of Antigravity 2.0's Dynamic Subagents capability, allowing these specialized agents to be spawned for parallel task execution without cluttering your main context. For example, the code-reviewer agent focuses solely on code quality and security.

Available Agents:

Skills

Define reusable workflows and domain knowledge. The coding-standards skill provides language-specific best practices.

Available Skills:

• coding-standards - Naming conventions, code structure, documentation standards
• backend-patterns - Backend architecture, API design, database patterns
• frontend-patterns - React/Next.js, state management, performance patterns
• tdd-workflow - Full TDD workflow with unit, integration, and E2E patterns
• security-review - Comprehensive security checklist and patterns
• verification-loop - Verification system for quality gates

Commands

Execute complex tasks with simple slash commands.

Native Antigravity 2.0 Commands (Built-in):

• /goal - Run a long-running task autonomously without stopping until the goal is achieved.
• /schedule - Set recurring crons or one-time timers for asynchronous automation.
• /grill-me - Align on a plan through an interactive interview.
• /browser - Launch a browser subagent to interact with web pages.
• /learn - Automatically extract patterns and persist behaviors.

Custom Workflow Commands (Included in this repo):

• /tdd - Test-driven development
• /plan - Implementation planning
• /code-review - Code review
• /build-fix - Incremental build error fixing
• /checkpoint - Workflow checkpoints
• /e2e - E2E test generation and execution
• /eval - Eval-driven development
• /learn - Extract reusable patterns as skills
• /orchestrate - Multi-agent workflow orchestration
• /refactor-clean - Safe dead code removal
• /test-coverage - Coverage analysis and test generation
• /update-codemaps - Architecture documentation
• /update-docs - Documentation sync
• /verify - Comprehensive codebase verification

Rules

Always-follow guidelines that apply to all code. Rules are automatically enforced by Antigravity.

Available Rules:

• security.md - Mandatory security checks (secrets, input validation, auth)
• coding-style.md - Code style and naming conventions
• git-workflow.md - Git workflow and commit guidelines
• testing.md - Testing requirements and TDD workflow
• patterns.md - Common reusable patterns
• performance.md - Performance optimization guidelines
• agents.md - Agent orchestration and model selection

Contexts

Dynamic system prompt injection for specific modes (development, research, review).

Available Contexts:

• dev.md - Active development mode
• research.md - Research and exploration mode
• review.md - Code review mode

MCP Servers

Model Context Protocol servers provide external integrations (GitHub, databases, APIs).

Configured Servers:

• github - GitHub operations (PRs, issues, repos)
• firecrawl - Web scraping and crawling
• supabase - PostgreSQL database operations
• memory - Persistent memory across sessions
• sequential-thinking - Chain-of-thought reasoning
• vercel - Vercel deployments and projects
• filesystem - Filesystem operations

Usage Examples

Planning a New Feature

User: I need to add real-time notifications when markets resolve

Antigravity: Let me create an implementation plan for you.
[Invokes planner agent]

Planner Agent:
# Implementation Plan: Real-Time Market Resolution Notifications

## Requirements
- Send notifications to users when markets they're watching resolve
- Support multiple channels (in-app, email, webhook)
- Ensure reliable delivery

## Implementation Phases

### Phase 1: Database Schema
- Add notifications table
- Add user_notification_preferences table
- Create indexes

### Phase 2: Notification Service
- Create notification service
- Implement queue using Redis
- Add retry logic

### Phase 3: Integration Points
- Hook into market resolution logic
- Query users with positions
- Enqueue notifications

### Phase 4: Frontend Components
- Create NotificationBell component
- Add NotificationList modal
- Implement real-time updates

## Risks
- HIGH: Email deliverability
- MEDIUM: Performance with 1000+ users per market

**WAITING FOR CONFIRMATION**: Proceed? (yes/no/modify)

Code Review

User: Review my changes

Antigravity: Running code review...
[Invokes code-reviewer agent]

Code Reviewer:
# Code Review Report

## Critical Issues (Must Fix)

### 1. Hardcoded API Key
**File:** src/api/client.ts:42
**Issue:** API key exposed in source code

```typescript
// ❌ Bad
const apiKey = "sk-abc123"

// ✅ Good
const apiKey = process.env.API_KEY

High Issues (Should Fix)

2. Missing Input Validation

File: src/api/markets.ts:15 Issue: User input not validated before database query

Recommendation: ❌ BLOCK - Fix critical issues before merging

### Test-Driven Development

User: /tdd Implement market search function

Antigravity: Starting TDD workflow... [Invokes tdd-guide agent]

TDD Guide:

Step 1: Write Failing Test (RED)

describe('searchMarkets', () => {
  it('returns markets matching query', async () => {
    const results = await searchMarkets('election')
    expect(results).toHaveLength(5)
  })
})

Running test... ❌ FAILED (expected - no implementation yet)

Step 2: Write Minimal Implementation (GREEN)

export async function searchMarkets(query: string) {
  // Minimal implementation
  return await db.markets.search(query).limit(5)
}

Running test... ✅ PASSED

Step 3: Refactor (IMPROVE)

[Suggests improvements while keeping tests green]

---

## Best Practices

### Agent Usage

1. **Use planner before coding** - Plan complex features before implementation
2. **Use code-reviewer after changes** - Review all code before committing
3. **Use tdd-guide for new features** - Write tests first, then implement
4. **Use security-reviewer for sensitive code** - Review auth, payments, user input
5. **Use build-error-resolver when stuck** - Fix TypeScript errors efficiently

### Skill Development

1. **Keep skills focused** - One skill per domain (backend, frontend, security)
2. **Update regularly** - Evolve skills as patterns emerge
3. **Document examples** - Include code examples in skills
4. **Version control** - Track skill changes in git

### Rule Management

1. **Keep rules minimal** - Only enforce critical guidelines
2. **Make rules specific** - Vague rules are hard to follow
3. **Provide examples** - Show good and bad patterns
4. **Review regularly** - Remove outdated rules

### MCP Configuration

1. **Don't enable all MCPs** - Keep under 10 enabled per project
2. **Use project-specific configs** - Different projects need different tools
3. **Secure your keys** - Never commit API keys to git
4. **Monitor context usage** - Too many MCPs reduce available context

---

## Customization

These configs are starting points. Customize for your stack:

### For Your Language/Framework

Create language-specific skills:

skills/ ├── python-conventions/ ├── react-patterns/ ├── go-best-practices/ └── rust-patterns/

### For Your Domain

Create domain-specific agents:

agents/ ├── ml-engineer.md # Machine learning specialist ├── devops-specialist.md # Infrastructure and deployment ├── data-engineer.md # Data pipelines and ETL └── mobile-developer.md # iOS/Android development

### For Your Team

Add team-specific rules to your `AGENTS.md` file:
```markdown
# AGENTS.md
- Team Coding Style
- Team Git Workflow
- Team Review Process
- Team Deployment

Contributing

Contributions are welcome! If you have useful agents, skills, or configurations for Antigravity, please submit a pull request.

Ideas for Contributions

• Language-specific skills (Python, Go, Rust patterns)
• Framework-specific configs (Django, Rails, Laravel)
• DevOps agents (Kubernetes, Terraform, AWS)
• Testing strategies (different frameworks)
• Domain-specific knowledge (ML, data engineering, mobile)

See CONTRIBUTING.md for guidelines.

Important Notes

Context Window Management

Critical: Don't enable all MCPs at once. Your context window can shrink significantly with too many tools enabled.

Rule of thumb:

• Have 20-30 MCPs configured
• Keep under 10 enabled per project
• Monitor context usage

Security

Never commit secrets:

• Use environment variables for API keys
• Add .env to .gitignore
• Use .env.example for documentation
• Rotate keys if accidentally committed

Performance

Optimize for speed:

• Use specific agents for specific tasks
• Keep skills focused and concise
• Disable unused MCP servers
• Monitor token usage

Comparison with Claude Code

This repository is inspired by everything-claude-code but adapted specifically for Google's Antigravity IDE.

Key Differences:

• Antigravity-specific agent configurations
• Gemini model references (gemini-2.0-pro-exp)
• Antigravity workflow patterns
• Context management for Antigravity

Shared Concepts:

• Agent-based delegation
• Skills for domain knowledge
• Rules for always-follow guidelines
• MCP server integrations

Resources

• Antigravity Documentation: [Link to Antigravity docs]
• MCP Protocol: https://modelcontextprotocol.io
• Inspired by: everything-claude-code

License

MIT - Use freely, modify as needed, contribute back if you can.

Built for the future of coding with Google Deepmind's Antigravity.

Star this repo if it helps. Customize for your needs. Build something great.