How to Contribute to Google Skills: A Complete Guide for the google/skills Repository

You can contribute to google/skills by reporting bugs, requesting new skills, or forking the repository to create your own custom skills—though the maintainers do not accept direct pull requests that modify existing skill definitions.

The google/skills repository is a curated collection of Agent Skills: step-by-step workflows for Google Cloud products and related technologies. While the core maintainers maintain strict control over the main branch, the project provides clear pathways for community involvement. Understanding these pathways and the repository's architecture will help you contribute effectively.

Understanding the Contribution Policy

Before attempting to contribute, review the official policy in CONTRIBUTING.md. The maintainers explicitly do not accept external pull requests that modify existing skill definitions. This policy preserves quality and consistency across the curated collection.

However, this restriction does not eliminate contribution opportunities. The project actively encourages:

• Bug reports and accuracy fixes through GitHub Issues
• Feature requests for new skills
• Community forks for customization and experimentation
• Documentation feedback via fork-based pull requests

Ways to Contribute to Google Skills

Report Bugs or Inaccuracies

The most direct way to improve the repository is identifying problems. Look for:

• Typos in skill descriptions or commands
• Outdated Google Cloud CLI commands or API references
• Broken examples that no longer execute correctly
• Missing prerequisites or incomplete step sequences

To report an issue, navigate to the Issues page and use the bug report template. Include the skill name, specific location of the problem, and reproduction steps when applicable.

Request New Skills

If you identify a gap in the Google Cloud workflow coverage, propose a new skill. The maintainers welcome feature requests that describe:

• A specific Google Cloud product pattern not currently documented
• Cross-product workflows combining multiple services
• Common operational tasks developers repeatedly perform

Submit these through the same Issues tracker using the feature request template.

Fork and Remix Skills

For hands-on contributors, forking provides full creative control. The repository's architecture supports this workflow explicitly.

# 1. Fork the repository on GitHub (click "Fork" button)

# 2. Clone your fork locally

git clone https://github.com/<your-username>/skills.git
cd skills

# 3. Create a new skill directory following the established pattern

mkdir -p skills/cloud/my-custom-skill
cat > skills/cloud/my-custom-skill/SKILL.md <<'EOF'

# My Custom Skill

## Description

A concise description of what this skill accomplishes.

## Steps

1. Execute the first operational step...
2. Execute the second operational step...
EOF

# 4. Commit and push to your fork

git add .
git commit -m "Add My Custom Skill"
git push origin main

After developing your skill, you can optionally link your fork in a GitHub Issue to share it with the maintainers and broader community.

Understanding the Google Skills Architecture

Contributing effectively requires understanding how skills are structured and discovered.

Skill File Organization

Each skill resides at skills/<category>/<skill-name>/SKILL.md. The SKILL.md file follows a strict specification defining:

• Metadata headers for discovery and installation
• Description of purpose and prerequisites
• Numbered steps for workflow execution
• Optional reference files for supporting content

Automatic Discovery Mechanism

The top-level README.md generates its "Available Skills" list by scanning the skills/ directory. Adding a valid SKILL.md to a new directory automatically makes the skill discoverable by the installation tools.

Installation and Plugin Integration

Users install skills via the official CLI:

npx skills add google/skills

# Select from the interactive UI, e.g., "Google Cloud Recipe: Foundation Builder"

The npx skills command reads SKILL.md metadata to resolve dependencies and present options. The repository also bundles plugins for AI agent platforms (Claude, Codex, Antigravity CLI) under plugins/, with lookup tables in the README connecting skills to agent integrations.

Key Files for Contributors

File	Purpose	Location
README.md	Entry point, skill listing, installation guide	Repository root
CONTRIBUTING.md	Official contribution policy and workflow	Repository root
SKILL.md (template)	Specification for skill structure	skills/<category>/<skill-name>/SKILL.md
skills.sh	Installer script referenced by CLIs	Referenced in documentation
plugins/	Agent platform integrations	plugins/ directory

Best Practices for Google Skills Contributions

When creating or modifying skills in your fork, follow these conventions:

• Use clear, action-oriented step descriptions—each step should be executable
• Include prerequisite sections specifying required permissions, tools, and knowledge
• Validate all commands against current Google Cloud documentation
• Test installation flow using npx skills add with your fork reference
• Reference the Apache 2.0 license in any distributed derivative works

Summary

• Direct PRs to main are restricted—review CONTRIBUTING.md before attempting changes
• GitHub Issues remain open for bugs, inaccuracies, and new skill requests
• Forking enables full customization while respecting the upstream curation model
• Skill structure is markdown-based—no complex build tooling required
• Discovery is automatic—properly placed SKILL.md files appear in installation UIs

Frequently Asked Questions

Can I submit a pull request directly to google/skills?

No. The maintainers do not accept external pull requests that modify existing skill definitions, as documented in CONTRIBUTING.md. Use GitHub Issues for bug reports and feature requests, or fork the repository for independent development.

How do I install a skill from my own fork?

Use the same npx skills CLI with your fork's GitHub reference: npx skills add <your-username>/skills. The interactive selector will present skills found in your fork's skills/ directory.

What license covers contributions to google/skills?

The repository uses the Apache 2.0 license, as specified in the LICENSE file. Any skills you create in a fork remain under this license, and you must maintain the license header when distributing derivative works.

Do I need Google Cloud expertise to contribute?

For meaningful skill development, yes—skills document operational workflows for Google Cloud products. However, you can contribute bug reports for documentation clarity without deep technical expertise, provided you verify the issue against the actual behavior described.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}