Google Antigravity Skills: Complete Guide [2026]
Everything you need to know about Google Antigravity agent skills — what they are, the best GitHub repos, 8 skill examples with templates, and how to create your own.
Quick Answer
Google Antigravity agent skills are SKILL.md files stored in .antigravity/skills/. They use the same format as Claude Code skills, making them fully portable. Find the best Antigravity skills on GitHub (search "antigravity SKILL.md") or browse the Guanyang community collection.
What is Google Antigravity?
Google Antigravity is an AI agent orchestration platform designed for coordinating multiple AI agents on complex software development tasks. Built to work with both Claude (Anthropic) and Gemini (Google), Antigravity excels at multi-step workflows that require parallel execution and agent handoffs.
Unlike single-agent IDEs, Antigravity routes subtasks to specialized agents — a planner agent, a coder agent, a reviewer agent — each powered by skills. This architecture makes agent skills even more valuable in Antigravity than in traditional single-agent setups.
Multi-Agent
Orchestrates teams of specialized AI agents working in parallel
Model Agnostic
Works with Claude, Gemini, and other models simultaneously
Skills-First
SKILL.md files are the primary way to define agent behaviors
For more on how Antigravity integrates with Claude and skills, see our guide to setting up AI skills in Claude Code, Gemini CLI, and Antigravity.
Antigravity Skills Format
Antigravity uses the standard SKILL.md format, with optional Antigravity-specific annotations for multi-agent coordination:
--- name: analyze-codebase description: > Perform deep architectural analysis of a codebase. Use when user wants to understand a new repo or plan a major refactor or migration. # Antigravity-specific (optional): agents: [planner, analyst] context: full-repo --- # Codebase Analyzer Skill ## Analysis Scope 1. Architecture patterns (MVC, layered, microservices, etc.) 2. Dependency graph (external packages, internal coupling) 3. Code quality metrics (complexity, duplication, coverage) 4. Security posture (auth, input handling, secrets) 5. Performance bottlenecks (hot paths, N+1s, memory) ## Output Format Executive summary (3-5 sentences) Architecture diagram (text-based) Top 5 risks ranked by severity Top 5 quick wins ranked by impact/effort Recommended next steps # Antigravity agent handoff: @handoff reviewer: "Review analysis for accuracy before presenting"
8 Best Google Antigravity Skills [2026]
The most effective Antigravity skills for development teams, with use cases and descriptions.
Codebase Analyzer
/analyze-codebaseDeep architectural analysis of a repository — identifies patterns, anti-patterns, dependencies, and improvement opportunities.
Use case: Onboarding to a new codebase or conducting architectural reviews before major refactors.
Test Suite Creator
/create-testsGenerates comprehensive test suites for any language — unit, integration, and e2e tests with mocked dependencies.
Use case: Adding test coverage to untested code or generating regression tests before a migration.
Query Optimizer
/optimize-queryAnalyzes SQL or NoSQL queries for performance, suggests indexes, rewrites for efficiency, and estimates impact.
Use case: Optimizing slow queries identified by monitoring tools or during database migrations.
Security Scanner
/security-scanPerforms OWASP-aligned security scan covering injection, auth, secrets, and dependency vulnerabilities.
Use case: Pre-release security review or continuous integration security gate.
Component Refactorer
/refactor-componentRefactors UI components to modern patterns — removes prop drilling, extracts hooks, improves accessibility.
Use case: Modernizing React class components or cleaning up components with too many responsibilities.
Documentation Generator
/generate-docsCreates comprehensive documentation from code — API docs, architecture guides, and README files.
Use case: Preparing open-source projects for contributors or documenting internal APIs for team onboarding.
Pipeline Debugger
/debug-pipelineAnalyzes CI/CD pipeline failures with step-by-step root cause analysis and fix recommendations.
Use case: When CI pipelines fail and the error messages are unclear or buried in logs.
Deploy Automation
/deploy-workflowCreates complete deployment workflows with pre-flight checks, blue-green strategies, and rollback procedures.
Use case: Setting up reliable deployment processes or migrating from manual to automated deployments.
Best Antigravity Skills Repos on GitHub
google/antigravity-awesome-skillsOfficial curated list of the best Antigravity skills — verified by the Antigravity team
guanyang/antigravity-skillsThe Guanyang collection — enterprise-grade skills for production Antigravity deployments
antigravity-community/skills-starterCommunity starter pack — 15 essential skills to bootstrap any Antigravity project
mlops/antigravity-data-skillsSpecialized skills for data engineering and ML pipeline workflows in Antigravity
How to Install Skills in Antigravity
# Method 1: Clone and copy git clone https://github.com/antigravity-community/skills-starter cp -r skills-starter/analyze-codebase .antigravity/skills/ # Method 2: Install from registry using skills.sh skills install analyze-codebase --platform antigravity # Method 3: Create from scratch mkdir -p .antigravity/skills/my-skill cat > .antigravity/skills/my-skill/SKILL.md << 'EOF' --- name: my-skill description: What this skill does and when to use it --- # My Skill [Instructions here] EOF # Trigger in Antigravity /my-skill
Create Antigravity Skills with the Skill Creator
Use the skill creator tool at mcpdirectory.app/create to generate SKILL.md files for Antigravity in minutes. Describe what you want the skill to do and the tool generates a complete, optimized SKILL.md ready to drop into your project.
✅ Tip: Skills created at mcpdirectory.app/create work identically in Antigravity, Claude Code, and Cursor — no modifications needed.
Frequently Asked Questions
What are Google Antigravity agent skills?
Google Antigravity is an AI agent orchestration platform that uses the SKILL.md format for reusable task instructions. Antigravity agent skills are SKILL.md files that define specific workflows for the Antigravity platform — the same format used in Claude Code and Cursor, making skills portable across all three platforms.
Where can I find Antigravity skills on GitHub?
Search GitHub for 'antigravity SKILL.md', browse the 'antigravity-skills' topic tag, or look for the google/antigravity-awesome-skills curated repository. The Guanyang collection is a well-known community skills library specifically for Antigravity workflows.
How do I install skills in Antigravity?
Create a .antigravity/skills/ directory in your project (or use ~/.antigravity/skills/ for global skills). Add your SKILL.md files there. Antigravity discovers skills on startup and makes them available as /commands in the agent chat. The process is identical to Claude Code's skill installation.
How do I create a skill for Google Antigravity?
Create a directory at .antigravity/skills/[skill-name]/ and write a SKILL.md file with YAML frontmatter (name, description) and your task instructions in markdown. The SKILL.md format is identical to Claude Code skills — if a skill works in Claude Code, it works in Antigravity.
What is the Antigravity skills documentation?
Antigravity uses the standard SKILL.md specification documented at mcpdirectory.app. The format is: YAML frontmatter with name and description fields, followed by markdown instructions. Antigravity-specific extensions include @context annotations for agent team coordination and @handoff directives for passing work between agents.
Build Your First Antigravity Skill
Use our skill creator to generate a complete Antigravity-compatible SKILL.md in under 60 seconds, or browse the curated skills catalog for ready-made options.