Skill Catalog & Data Management

Skill Catalog & Data Management

Antigravity Awesome Skills is built upon a robust system for managing, indexing, and distributing a growing library of agentic skills. This section details how skills are cataloged, their structure, and the data management processes that ensure users have access to a searchable and up-to-date collection.

The Awesome Skills Catalog

The core of Antigravity Awesome Skills is its comprehensive catalog, which serves as the central index for all available agentic skills. This catalog enables discovery, browsing, and installation of skills across various AI coding assistants.

skills.json: The Public-Facing Catalog

For users and the web application, the primary interface to the skill data is often through a generated skills.json file. This file contains a consolidated list of all skills, complete with their essential metadata. It's optimized for quick loading and filtering, powering the searchable skill browser and enabling tools to quickly reference available skills without needing to parse individual skill directories.

Each entry in skills.json typically includes:

• id: A unique identifier for the skill.
• title: A human-readable name for the skill.
• description: A brief summary of what the skill does.
• category: The primary classification of the skill (e.g., coding, debugging, security).
• tools: An array of compatible AI assistants (e.g., claude-code, cursor, gemini-cli).
• tags: Keywords for further filtering and search.
• path: The relative path to the skill's directory within the repository.
• version: The version of the skill, if applicable.

This structured metadata allows agents and applications to programmatically interact with the skill library, improving the contextual understanding and execution of tasks.

Internal Indexing: catalog.json and skills_index.json

Underpinning the public skills.json are internal data artifacts like catalog.json and skills_index.json. These files represent intermediate or comprehensive views of the skill data used during the build and synchronization processes. They might contain more detailed or raw information about each skill, including file paths, asset lists, and historical metadata, before being processed and optimized into the user-facing skills.json. While these files are internal to the project's data management pipeline, they are critical components in assembling and maintaining the integrity of the overall skill catalog.

Skill Structure

Each "skill" in Antigravity Awesome Skills is typically a self-contained unit, often residing in its own directory. A standard skill structure includes:

• SKILL.md: The primary playbook document. This Markdown file contains the structured instructions, context, constraints, and output requirements that an AI agent uses to perform the task. It's designed to be human-readable and machine-interpretable.
• assets/: A directory containing supporting files such as:
  • Code templates (e.g., rest-api-template.py).
  • Scripts (e.g., api_validator.py, api_bench.py).
  • Configuration files.
  • Example inputs/outputs.
• metadata.json: (Implicit or sometimes explicit) A file providing additional machine-readable metadata specific to that skill, extending what's in the main catalog.

This modular structure ensures that skills are portable, maintainable, and easy for both humans and AI agents to understand and utilize.

Metadata Synchronization and Updates

The antigravity-awesome-skills repository is continuously updated. The <!-- registry-sync --> comment in the README.md indicates an automated synchronization process that regularly scans the repository, updates skill counts, and regenerates the catalog files. This process ensures that new skills, updates to existing skills, and changes in metadata are promptly reflected in the public catalog, providing users with the most current and accurate information.