46 Essential Terms

AI Skills Glossary

Master the language of AI-powered development. 46 quotable definitions covering skills, commands, agents, and the Claude Code ecosystem.

Last updated: June 2026

Core Concepts

Fundamental concepts of the AI skills ecosystem and Claude Code marketplace.

AI Skill: A packaged capability that extends Claude Code with new commands, agents, or tools. Skills are installed from the marketplace and immediately available in your CLI.
Claude Code: The official CLI tool from Anthropic that brings Claude AI into your terminal. It serves as the runtime environment for skills, commands, and AI-powered development workflows.
MCP (Model Context Protocol): An open protocol developed by Anthropic that enables AI assistants to securely connect to external data sources and tools. MCP allows Claude to access databases, APIs, and file systems through standardized interfaces.
AI Skill Market: A curated marketplace where developers discover, install, and share production-ready AI capabilities for Claude Code. Think npm for AI skills.
Skill Publisher: A developer or organization that creates and distributes skills on the AI Skill Market. Publishers can be verified to indicate trustworthiness and quality.
Installation: The process of adding a skill to your Claude Code environment, typically accomplished by downloading the skill folder to ~/.claude/skills/. Most skills install in under 60 seconds.
Slash Command: A command invoked with a forward slash (e.g., /commit, /test) that executes a specific skill or action. Slash commands provide quick access to frequently-used workflows.
Agent: An autonomous AI workflow that breaks complex tasks into steps, using reasoning and tool-calling to achieve goals without constant human intervention.
Skill Registry: The centralized database that tracks all available skills, their metadata, install counts, and quality scores. The registry powers search and discovery in the marketplace.
Workflow: A multi-step process automated by a skill or agent, such as "test-driven development" or "pull request creation." Workflows chain together multiple commands and decisions.

Skill Types

Different categories of skills you can install and use with Claude Code.

Command Skill: A skill that adds executable slash commands to Claude Code (e.g., /commit, /deploy). Command skills are triggered explicitly by the user and perform specific, well-defined actions.
Agent Skill: An autonomous skill that operates independently to complete complex tasks by breaking them into subtasks. Agent skills use reasoning, planning, and tool-calling to achieve goals.
MCP Skill: A skill that implements the Model Context Protocol to connect Claude to external services like databases, APIs, or file systems. MCP skills enable Claude to read and write data beyond the local filesystem.
Workflow Skill: A skill that orchestrates multi-step processes involving multiple tools and decisions. Workflow skills automate repetitive development patterns like feature branches, testing, and deployment.
Integration Skill: A skill that connects Claude Code to third-party services like GitHub, Jira, Slack, or cloud providers. Integration skills enable seamless cross-platform workflows.
Analysis Skill: A skill that examines codebases, logs, or data to provide insights, detect issues, or generate reports. Analysis skills help with code review, performance profiling, and security auditing.
Generation Skill: A skill that creates new code, documentation, tests, or configuration files based on templates or AI reasoning. Generation skills accelerate boilerplate creation and scaffolding.
Automation Skill: A skill that eliminates manual tasks by automating repetitive workflows like dependency updates, changelog generation, or release tagging.
Utility Skill: A general-purpose skill providing helper functions, formatters, or converters. Utility skills handle common tasks like JSON formatting, base64 encoding, or timestamp conversion.
Template Skill: A skill that scaffolds new projects, components, or files from pre-defined templates. Template skills enforce consistency and best practices across codebases.
Plugin System: The extensibility architecture that allows developers to add custom functionality to Claude Code through skills. The plugin system manages skill discovery, loading, and lifecycle.
Composite Skill: A skill that combines multiple other skills or commands into a unified workflow. Composite skills create powerful automation by chaining simpler capabilities.

Quality Metrics

How we measure, verify, and rank skills in the marketplace.

Quality Score: A 1-5 star rating that measures skill reliability based on documentation completeness, test coverage, install success rate, and user feedback. Skills rated 4+ stars receive "High Quality" badges.
Install Count: The total number of times a skill has been installed by developers worldwide. High install counts indicate popularity and community trust.
Verified Publisher: A skill creator whose identity and code quality have been validated by the AI Skill Market team. Verified publishers display a blue checkmark badge.
Official Skill: A skill published directly by Anthropic or from official Anthropic repositories. Official skills display a purple badge and represent the highest quality standard.
Popular Badge: An indicator awarded to skills with 100+ installations, signaling widespread adoption and community validation.
High Quality Badge: An indicator awarded to skills with quality scores of 4 stars or higher, denoting excellent documentation, testing, and reliability.
Trending Skill: A skill experiencing rapid growth in installations (50+ installs within 90 days). Trending skills highlight emerging capabilities gaining community traction.
New Badge: An indicator for skills published within the last 30 days, helping users discover fresh capabilities recently added to the marketplace.
Success Rate: The percentage of skill installations that complete without errors. High success rates (95%+) indicate robust dependency management and compatibility.

Technical Terms

Technical concepts, architecture patterns, and development terminology.

Server Component: A Next.js React component that renders exclusively on the server, enabling direct database access and optimal performance. Server Components fetch data at build time or request time.
Server Action: A Next.js pattern for handling form submissions and data mutations on the server without API routes. Server Actions use the "use server" directive and run in secure server environments.
Row Level Security (RLS): Database-level access control in Supabase that enforces permissions at the row level. RLS policies ensure users only access data they are authorized to view or modify.
Realtime Subscription: A Supabase feature that streams database changes to clients in real-time using WebSockets. Realtime subscriptions power live install counts and dynamic skill updates.
JSON-LD: A structured data format (JSON for Linking Data) that helps search engines understand page content and display rich results. JSON-LD enables features like knowledge panels and FAQ snippets.
Structured Data: Machine-readable information about web page content following Schema.org standards. Structured data improves SEO by enabling rich search results and better content understanding.
Metadata API: Next.js built-in API for defining page titles, descriptions, and social sharing tags. The Metadata API supports static and dynamic metadata generation for optimal SEO.
Static Generation: A Next.js rendering strategy that pre-builds pages at build time, delivering HTML instantly without server processing. Static generation offers the best performance for content-heavy pages.
Client Component: A React component that runs in the browser, enabling interactivity through hooks like useState and useEffect. Client Components are marked with the "use client" directive.
Sandboxing: A security mechanism in Claude Code that restricts skill access to filesystem, network, and system resources. Sandboxing prevents malicious skills from compromising user systems.
Environment Variable: A dynamic configuration value stored outside source code (e.g., API keys, database URLs). Environment variables keep secrets secure and enable different configurations per environment.
API Route: A serverless function in Next.js that handles HTTP requests at /api/* paths. API routes enable backend logic without a separate server.
Webhook: An HTTP callback that sends real-time notifications when events occur. Webhooks enable skills to react to external events like GitHub pushes or deployment completions.
Shadcn UI: A collection of reusable React components built with Radix UI and Tailwind CSS. Shadcn UI provides accessible, customizable components that you copy into your project rather than installing as a dependency.
Tailwind CSS: A utility-first CSS framework that enables rapid UI development through composable class names. Tailwind eliminates custom CSS files by providing low-level utility classes for styling.

Ready to explore AI Skills?

Browse 4,600+ production-ready skills and start shipping AI capabilities 60X faster.

Browse Skills