Where Are Individual Skills Located in the Google Skills Repository?

Individual skills in the google/skills repository (also known as Instagit) are stored as SKILL.md files within a hierarchical directory structure following the pattern skills/<domain>/<skill-name>/SKILL.md.

The google/skills repository organizes technical documentation and reference materials as discrete, reusable units called skills. Understanding the exact location of individual skills within the repository enables developers to browse documentation, construct direct GitHub links, and programmatically index available resources.

Hierarchical Directory Structure

Individual skills follow a strict three-tier directory convention. At the repository root, the skills/ directory contains domain-specific subdirectories (such as cloud, analytics, ads, and developers). Within each domain folder, individual skill directories use kebab-case naming conventions and contain exactly one SKILL.md file that serves as the canonical definition for that skill.

The repository layout follows this pattern:

skills/
├─ analytics/
│   ├─ google-analytics-data-api-basics/
│   │   └─ SKILL.md
│   └─ google-analytics-admin-api-basics/
│       └─ SKILL.md
├─ ads/
│   ├─ ima-sdk-client-side/
│   │   └─ SKILL.md
│   └─ google-mobile-ads-rewarded/
│       └─ SKILL.md
├─ cloud/
│   ├─ workload-manager-basics/
│   │   └─ SKILL.md
│   ├─ spanner-basics/
│   │   └─ SKILL.md
│   └─ gke-basics/
│       └─ SKILL.md
└─ developers/
    ├─ retrieving-developer-knowledge/
    │   └─ SKILL.md
    └─ finding-google-skills/
        └─ SKILL.md

All skills adhere to the path structure: skills/<domain>/<skill-name>/SKILL.md.

Domain Organization

Skills are grouped into logical domains to categorize related functionality:

• cloud: Google Cloud Platform services including Workload Manager, Spanner, and GKE
• analytics: Google Analytics Data API and Admin API documentation
• ads: Interactive Media Ads (IMA) SDK and mobile advertising implementations
• developers: Developer productivity tools and knowledge retrieval systems

Each domain directory contains multiple skill subdirectories. For example, skills/cloud/ houses workload-manager-basics/, spanner-basics/, and gke-basics/, with each subdirectory containing its respective SKILL.md file.

Programmatically Locating Skills

Developers can discover and access individual skills using command-line utilities or Python scripts to navigate the repository structure without manual browsing.

Using Command-Line Tools

To enumerate all available skills, use the find command to locate every SKILL.md file:

# List all available skills (Linux/macOS)

find skills -type f -name SKILL.md

To inspect the contents of a specific skill, reference the full path relative to the repository root:

# Display the first 20 lines of the GKE basics skill

cat skills/cloud/gke-basics/SKILL.md | head -n 20

Using Python for Discovery

For automated tooling or integration scripts, use Python's pathlib module to programmatically discover all skill definitions:

import pathlib

# Programmatically discover all skill definitions

repo_root = pathlib.Path(__file__).parent.parent
skill_files = list(repo_root.glob("skills/**/SKILL.md"))
for f in skill_files:
    print(f.relative_to(repo_root))

This returns all paths matching the skills/**/SKILL.md pattern, enabling dynamic indexing of the repository's documentation.

Direct GitHub Access

Individual skills are directly browsable on GitHub without cloning the repository. Construct URLs using the pattern:

https://github.com/google/skills/blob/main/skills/<domain>/<skill-name>/SKILL.md

Specific examples include:

• Workload Manager basics: skills/cloud/workload-manager-basics/SKILL.md
• Google Analytics Data API basics: skills/analytics/google-analytics-data-api-basics/SKILL.md
• IMA SDK client-side: skills/ads/ima-sdk-client-side/SKILL.md

Summary

• Individual skills are located under the skills/ directory following the strict pattern skills/<domain>/<skill-name>/SKILL.md
• SKILL.md files serve as the single source of truth and core definition for each skill
• Skills are organized by domain into cloud, analytics, ads, and developers categories
• Skills can be discovered using standard Unix tools (find, cat) or Python's pathlib module with glob patterns
• Each skill directory uses kebab-case naming and contains exactly one SKILL.md file

Frequently Asked Questions

What is the file naming convention for individual skills?

Every individual skill is defined by a single file named exactly SKILL.md located within its own subdirectory. The parent directory uses kebab-case naming (lowercase words separated by hyphens) to describe the skill's function, such as google-analytics-data-api-basics or retrieving-developer-knowledge.

How are skills organized within the repository structure?

Skills are grouped by domain into subdirectories immediately under skills/. Current domains include cloud for Google Cloud Platform services, analytics for measurement APIs, ads for advertising SDKs, and developers for engineering resources. This hierarchical organization ensures related skills are co-located for easier discovery.

Can I access individual skills directly through GitHub's web interface?

Yes, all skills are immediately accessible via GitHub using the URL pattern https://github.com/google/skills/blob/main/skills/<domain>/<skill-name>/SKILL.md. You can navigate the repository folder tree or construct direct links to specific skills without cloning the repository locally.

How can I programmatically list all available skills in the repository?

Use the find command to locate all skill definitions: find skills -type f -name SKILL.md. Alternatively, use Python's pathlib with the glob pattern skills/**/SKILL.md to generate a list of all skill paths programmatically, enabling automated documentation indexing or validation workflows.

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