How to Discover Available Skills in the google/skills Repository

The google/skills repository maintains a generated catalog file named index.json at the repository root that contains every available skill with its name, description, and entrypoint URL.

The google/skills repository organizes self-contained guides and automations as discrete "skills" scattered throughout its directory structure. To discover available skills programmatically without manually browsing directories, you interact with the central catalog file and understand its schema. This guide explains how to access, parse, and utilize the skill discovery mechanisms implemented in the repository.

The Central Skill Catalog (index.json)

At the root of the repository, the file index.json serves as the machine-readable index of all publishable skills. This JSON file is automatically regenerated on every push to ensure it reflects the current state of the repository.

Each entry in the skills array contains three critical fields:

• name – The short identifier used in prompts (e.g., gke-basics).
• description – A concise summary explaining the skill's purpose and scope.
• entrypoint – A URL pointing to the full markup documentation of the skill (SKILL.md).

Because the catalog is approximately 75KB in size, fetching and parsing it incurs minimal performance overhead.

Step-by-Step Discovery Workflow

The standard process for discovering skills follows four steps:

1. Fetch the catalog – Retrieve the raw JSON from https://raw.githubusercontent.com/google/skills/main/index.json.
2. Parse the JSON – Extract the skills array containing all skill records.
3. Enumerate entries – Filter or display the name and description fields as needed.
4. Load specific skills – Use the entrypoint URL to fetch the corresponding SKILL.md file for full instructions.

Programmatic Approaches to Access the Catalog

You can implement this workflow using various programming languages and tools to discover available skills in the google/skills repository.

Shell with curl and jq

For command-line environments, combine curl with jq to download and filter the catalog:

# Download the catalog and list every skill name with its description

curl -sS https://raw.githubusercontent.com/google/skills/main/index.json | \
  jq -r '.skills[] | "\(.name)\t\(.description)"'

Python with requests

Use the requests library to fetch and iterate over the catalog:

import requests
import json

url = "https://raw.githubusercontent.com/google/skills/main/index.json"
resp = requests.get(url)
catalog = resp.json()

for skill in catalog["skills"]:
    print(f"{skill['name']}: {skill['description']}")

Node.js with fetch

In Node.js environments, use the fetch API (or node-fetch for older versions):

const fetch = require('node-fetch');

(async () => {
  const res = await fetch('https://raw.githubusercontent.com/google/skills/main/index.json');
  const data = await res.json();

  data.skills.forEach(s => {
    console.log(`${s.name}: ${s.description}`);
  });
})();

Leveraging the Google Skill Finder Helper

The repository includes a specialized skill called Google Skill Finder (finding-google-skills) located at skills/developers/finding-google-skills/SKILL.md. If you operate within a Google-Skills-enabled environment, invoke this skill at the START of any request involving Google products.

This helper skill automatically fetches the catalog, matches your request against skill descriptions, and returns the appropriate entrypoint URLs. The full workflow and implementation details are documented in its source file at skills/developers/finding-google-skills/SKILL.md.

Key Files for Skill Discovery

Understanding the repository structure helps contextualize the discovery process:

• index.json – The generated catalog in the repository root containing all skill metadata.
• skills/developers/finding-google-skills/SKILL.md – Documentation for the helper skill that automates discovery.
• README.md – High-level repository overview explaining skill organization.
• skills/**/SKILL.md – Individual skill definitions referenced by catalog entrypoint fields.

Summary

• The google/skills repository uses a generated index.json file at the repository root as the canonical catalog of all skills.
• Each catalog entry provides a name, description, and entrypoint URL pointing to the skill's SKILL.md file.
• You can discover skills programmatically by fetching and parsing this JSON file using shell commands, Python, or Node.js.
• The finding-google-skills skill provides an automated discovery mechanism for Google-Skills-enabled environments.
• The catalog regenerates automatically on every push, ensuring it always reflects the current available skills.

Frequently Asked Questions

Where is the skill catalog located in the google/skills repository?

The skill catalog is located at the repository root in the file index.json. This file is automatically generated from all SKILL.md files scattered throughout the repository structure.

What information does each entry in the skill catalog contain?

Each entry contains three fields: name (the identifier used in prompts), description (a summary of the skill's purpose), and entrypoint (a URL pointing to the full skill documentation).

How do I access the full documentation for a specific skill after discovering it?

Use the entrypoint URL provided in the catalog entry. This URL points directly to the skill's SKILL.md file, which you can fetch to access complete instructions and implementation details.

Is the skill catalog updated automatically?

Yes. According to the repository implementation, index.json regenerates automatically on every push, ensuring the catalog always reflects the current set of available skills without manual intervention.

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