gemini-extension-development

Gemini Extension Development

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "gemini-extension-development" with this command: npx skills add melodic-software/claude-code-plugins/melodic-software-claude-code-plugins-gemini-extension-development

Gemini Extension Development

Documentation Delegation

Documentation Source: For authoritative extension syntax and current features, query gemini-cli-docs skill. This skill provides development guidance; gemini-cli-docs provides official Gemini CLI documentation.

Overview

Expert skill for creating, testing, and distributing Gemini CLI Extensions. Extensions are the primary plugin mechanism for Gemini.

When to Use This Skill

Keywords: gemini extension, create extension, gemini extensions link, extension gallery, context files, extension commands, extension releasing, uninstall extension

Use this skill when:

  • Creating a new extension (gemini extensions create )

  • Linking a local extension for development (gemini extensions link )

  • Packaging MCP servers into extensions

  • Adding custom slash commands (.toml ) to extensions

  • Installing extensions from GitHub or local paths

  • Releasing extensions via Git or GitHub Releases

Extension Anatomy

An extension can contain:

  • extension.yaml : Manifest file.

  • GEMINI.md : Context "playbook" for the model.

  • package.json : Dependencies (if using Node.js/TypeScript). Note: Use the Unified Google Gen AI SDK (e.g., google-genai ) as google-generativeai is deprecated.

  • MCP Servers: Embedded tools.

  • Commands: *.toml files defining custom slash commands.

  • Tool Restrictions: excludeTools configuration.

Development Workflow

  • Create: gemini extensions create my-extension

  • Link: cd my-extension && gemini extensions link . (Enables hot-reloading)

  • Test: Run gemini and use the new capabilities.

  • Publish: Push to GitHub (installable via URL).

Keyword Registry (Delegates to gemini-cli-docs)

Topic Query Keywords

Creation create extension , extension template

Manifest extension.yaml schema , extension manifest

Commands extension slash commands , toml commands

Linking gemini extensions link , local extension dev

Releasing extension releasing git , github release extension

Management uninstall extension , update extension

Quick Decision Tree

What do you want to do?

  • Start a New Extension -> Query gemini-cli-docs : "create extension boilerplate"

  • Test Locally -> Query gemini-cli-docs : "link local extension"

  • Add a Command -> Query gemini-cli-docs : "define command in extension"

  • Bundle an MCP Server -> Query gemini-cli-docs : "extension mcp server"

  • Install an Extension -> Query gemini-cli-docs : "install extension from url"

  • Release an Extension -> Query gemini-cli-docs : "extension releasing git vs github"

Test Scenarios

Scenario 1: Create Extension

Query: "How do I create a new Gemini CLI extension?" Expected Behavior:

  • Skill activates on "create extension"

  • Delegates to gemini-cli-docs for template command Success Criteria: User receives gemini extensions create syntax

Scenario 2: Link for Development

Query: "How do I test my Gemini extension locally?" Expected Behavior:

  • Skill activates on "test extension" or "link extension"

  • Provides gemini extensions link . workflow Success Criteria: User receives local development workflow

Scenario 3: Release Extension

Query: "How do I publish my Gemini extension?" Expected Behavior:

  • Skill activates on "release extension" or "publish"

  • Delegates to docs for Git/GitHub release options Success Criteria: User receives release workflow options

References

Official Documentation: Query gemini-cli-docs for:

  • "extensions"

  • "extension development"

Version History

  • v1.1.0 (2025-12-01): Added Test Scenarios section

  • v1.0.0 (2025-11-25): Initial release

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Coding

design-thinking

No summary provided by upstream source.

Repository SourceNeeds Review
Coding

plantuml-syntax

No summary provided by upstream source.

Repository SourceNeeds Review
Coding

system-prompt-engineering

No summary provided by upstream source.

Repository SourceNeeds Review