A curated list of LLM Skills, this repository provides resources and tools for customizing AI agent workflows. It targets developers and power users aiming to enhance LLM capabilities across platforms like Anthropic's Claude Code/Desktop, Google's Gemini CLI, and open-source alternatives. The core benefit is enabling LLMs to execute specific, repeatable tasks in a standardized manner, improving reliability and efficiency.

How It Works

LLM Skills are defined as modular, reusable workflows, each typically including a SKILL.md file. This markdown document outlines the skill's name, description, intended use cases, detailed instructions for the LLM, and practical examples. Platforms such as Claude Code automatically discover skills placed in designated project-local or user directories. For other command-line interfaces (CLIs), skills are integrated by referencing their SKILL.md content within prompts, often using file attachment mechanisms, allowing LLMs to follow structured operational guidance.

Quick Start & Requirements

• Claude Code: Place skill folders within .claude/skills/ (project-local) or ~/.claude/skills/ (user-level). Requires Node 20+ and Visual Studio Code.
• Claude Desktop: Upload skills as ZIP archives via Settings → Capabilities → Skills.
• Gemini CLI: Reference SKILL.md in prompts; supports @ file/context attachments for skill content. Requires Node 20+.
• OpenCode: Place skills folder in the project and instruct OpenCode to read the SKILL.md file.
• General: Installation for specific CLIs varies (e.g., npm install -g @google/gemini-cli, curl -fsSL https://opencode.ai/install | bash).

Highlighted Details

• Features a wide array of skills spanning Notion integration, document processing (docx, pdf, xlsx), web development (Playwright, AWS CDK), data analysis (CSV, PostgreSQL), creative media generation (imagen, Slack GIF), and security forensics.
• Skills emphasize specific, repeatable tasks, providing clear instructions, examples, and guidance on error handling for LLMs.
• Facilitates customization across diverse LLM platforms, including proprietary services (Claude, Gemini) and open-source CLIs.

Maintenance & Community

• Official documentation includes an LLM Skills Overview, User Guide, Development Guide, API Documentation, and related blog posts.
• Community resources comprise the Anthropic Skills Repository, Claude Community forums for discussion, and a Skills Marketplace for discovery and sharing.

Licensing & Compatibility

• The repository itself is licensed under the Apache License 2.0.
• Individual skills may carry different licenses; users must verify the specific license within each skill's directory. Commercial use compatibility is contingent on these individual licenses.

Limitations & Caveats

• CLIs such as Codex, Gemini CLI, OpenCode, and Qwen Code lack native support for Anthropic's Skills format, necessitating manual adaptation by incorporating SKILL.md content directly into prompts.
• Setup procedures and dependencies differ significantly across platforms, potentially requiring specific Node.js versions or command-line tool installations.