ANTIGRAVITY LABJP
Articles/Agents & Manager
Agents & Manager/2026-03-16Beginner

skills.sh Skill Agent Installation Guide

Complete guide to installing, customizing, and leveraging the skills.sh marketplace within Antigravity to automate development workflows with AI agent skills

skills.shClaude Code8agent skills2skill marketplaceworkflow automation2development efficiency

Setup and context

One of the most powerful features of Antigravity (Claude Code & Cowork) is skills.sh, an open skill marketplace for shared AI agent capabilities. This guide teaches you how to install, customize, and leverage skills.sh to automate your development workflows.

What is skills.sh?

skills.sh is an open-source AI agent skill marketplace where developers share reusable automation tasks. Key features include:

  • Open Source Management: Community-created skills published and versioned
  • Broad Functionality: Frontend design, SEO audits, PDF processing, data analysis, and more
  • Reusable Components: Build on existing skills or customize for your needs
  • Explicit Versioning: Clear dependency management and version tracking

Skill Structure

Each skill consists of:

  • SKILL.md: Metadata and description
  • Execution Scripts: Python, JavaScript, Bash, etc.
  • Dependency Files: requirements.txt, package.json, etc.
  • Test Cases: Verification and validation tests

Installation Methods

Basic Installation

claude install skills.sh/skill-name

For example, to install the frontend-design skill:

claude install skills.sh/frontend-design

Check Installed Skills

claude skills list

Install Specific Version

claude install skills.sh/skill-name@1.2.3

Popular Skills and Use Cases

1. frontend-design Skill

Purpose: UI/UX design generation, responsive layout analysis

claude install skills.sh/frontend-design

Use Cases:

  • Auto-generate responsive design proposals
  • Accessibility compliance checking
  • Color scheme optimization

2. seo-audit Skill

Purpose: SEO analysis, keyword research, meta tag optimization

claude install skills.sh/seo-audit

Use Cases:

  • Calculate SEO scores for blog posts
  • Competitive site analysis
  • Internal link structure improvements

3. pdf Skill

Purpose: PDF generation, parsing, and manipulation

claude install skills.sh/pdf

Use Cases:

  • Merge Markdown files into PDF
  • Auto-generate reports
  • Extract and organize text from PDFs

4. xlsx Skill

Purpose: Excel file generation and analysis

claude install skills.sh/xlsx

Use Cases:

  • Auto-generate data analysis reports
  • Convert and format CSV to Excel
  • Create multi-sheet workbooks

5. skill-creator Skill

Purpose: Scaffold new skills from templates

claude install skills.sh/skill-creator

Use Cases:

  • Accelerate team-specific skill development
  • Generate consistent skill scaffolding
  • Automate documentation creation

Creating Custom Skills

Step 1: Generate Skill Scaffolding

claude invoke skill-creator --name "my-custom-skill" --language "python"

Generated structure:

my-custom-skill/
├── SKILL.md           # Skill description
├── main.py            # Main script
├── requirements.txt   # Python dependencies
├── tests/
│   └── test_main.py   # Test cases
└── README.md          # Usage documentation

Step 2: Implement Skill Logic

Edit main.py:

def execute(input_data):
    """Main skill execution"""
    result = process_input(input_data)
    return {
        "status": "success",
        "output": result
    }
 
def process_input(data):
    # Custom processing
    return data.upper()

Step 3: Specify Dependencies

Update requirements.txt:

requests>=2.28.0
beautifulsoup4>=4.11.0

Step 4: Local Testing

cd my-custom-skill
pytest tests/

Step 5: Publish to skills.sh

claude publish my-custom-skill --registry skills.sh

Workflow Automation with Skill Composition

Combine multiple skills to create powerful automation pipelines.

Example 1: Blog Article Pipeline

  1. content-generator creates initial article
  2. seo-audit optimizes for search
  3. pdf generates report
claude workflow create blog-pipeline \
  --steps "content-generator,seo-audit,pdf"

Example 2: Data Analysis Pipeline

  1. data-import loads CSV files
  2. analytics performs analysis
  3. xlsx exports formatted results
claude workflow create analytics-pipeline \
  --steps "data-import,analytics,xlsx"

Run Workflow

claude workflow run blog-pipeline --input "topic=Python development"

Team Skill Sharing and Management

Create Private Skill Registry

claude config set skill-registry https://internal.skills.company.com

Manage Skill Permissions

claude skill grant --name "analytics-skill" \
  --team "data-team" --permission "execute"

Version Locking

# Pin specific version as default
claude skill pin "analytics-skill@2.1.0" --team "data-team"

Team Documentation

Include team-specific setup in README:

## Team Configuration
 
### Required Environment Variables
- API_KEY: Analytics API key
- DATABASE_URL: Database connection string
 
### Usage Example
```bash
claude invoke analytics-skill --param "mode=production"

## Troubleshooting

### Version Incompatibility

**Symptom**: `Incompatible skill version`

**Solution**:

```bash
# Check installed version
claude skill info skill-name

# Upgrade
claude upgrade skill-name

Missing Dependencies

Symptom: Missing required dependency

Solution:

# Reinstall with dependencies
claude install skill-name --reinstall-deps
 
# Or manually install
pip install -r requirements.txt

Skill Execution Errors

Symptom: Errors during skill execution

Debug Steps:

  1. Check skill logs:
claude logs skill-name
  1. Validate skill:
claude skill validate skill-name
  1. Run tests:
cd skill-name
pytest tests/ -v

Permission Errors

Symptom: Permission denied

Solution:

# Check skill directory permissions
ls -la ~/.claude/skills/skill-name/
 
# Fix permissions
chmod +x ~/.claude/skills/skill-name/main.py

Best Practices

Skill Development

  1. Clear Input/Output: Document expected formats thoroughly
  2. Error Handling: Handle unexpected inputs gracefully
  3. Test Coverage: Aim for 80%+ code coverage
  4. Performance: Measure and optimize execution time

Skill Usage

  1. Version Locking: Lock versions in production
  2. Dry Runs: Test before production runs
  3. Audit Logging: Track all skill executions
  4. Update Strategy: Regularly check for new versions

Summary

Leveraging skills.sh enables:

  • Significant improvement in development efficiency
  • Simplified implementation of complex workflows
  • Knowledge sharing across teams
  • Robust automation in production environments

Start exploring the Antigravity ecosystem today!


Related Resources

Share

Thank You for Reading

Antigravity Lab is ad-free, supported entirely by members like you. We publish practical guides daily with implementation code, benchmarks, and production-ready patterns. If you've found it useful, we'd love to have you on board.

  • Copy-paste ready implementation code
  • New advanced guides published daily
  • $5/mo or $10 for lifetime access
View Membership →

If you found this article helpful, a small tip ($1.50) would mean a lot to us. Your support helps keep this site ad-free and covers server and hosting costs.

Related Articles

Agents & Manager2026-05-06
gh skill: Sharing AI Agent Knowledge Across Claude Code, Copilot, Cursor, and Gemini CLI
The gh skill GitHub CLI extension lets you package SKILL.md definitions and distribute them across 30+ AI coding agents. Here's how it works and how to get started.
Agents & Manager2026-05-05
2026 AI Coding Agents Compared: How to Use Claude Code, OpenCode, Cursor, and Gemini CLI Together
The 2026 AI coding agent landscape is fragmented on purpose. Claude Code, OpenCode, Cursor, and Gemini CLI each have distinct strengths. This guide builds a practical framework for choosing and combining them based on project needs.
Agents & Manager2026-03-15
Google ADK × Antigravity: Build Custom Agent Skills to Extend Your AI
Learn how to build custom agent skills for Antigravity using Google's Agent Development Kit (ADK). From writing SKILL.md instructions to implementing scripts and deploying a real GitHub Issues triage skill — step by step.
📚RECOMMENDED BOOKS
Build a Large Language Model (From Scratch)
Sebastian Raschka
LLM Dev
Prompt Engineering for LLMs
Berryman & Ziegler
Prompting
AI Engineering
Chip Huyen
AI Eng
* Contains affiliate links
See all →