8 Must-Collect Agent Skills Repositories for Claude and AI Agents

Agent Skills are folders that contain a SKILL.md file with metadata, instructions, and optional helper scripts or reference documents. The file defines how an AI agent should perform a specific task.

my-skill/
├── SKILL.md   # Required: instructions + metadata
├── scripts/   # Optional: helper scripts
└── references/ # Optional: docs and examples

The SKILL.md file has two sections: a YAML front‑matter with metadata such as name, description, and triggers, and a markdown body containing the actual commands. When an agent executes a task, it scans the available skills, selects those whose triggers match the task, and loads their context.

Creating a skill from scratch can require days of work to understand tools, write clear prompts, test edge cases, and polish metadata. A curated skill library therefore provides ready‑to‑use, production‑grade capabilities.

Awesome Claude Skills (Composio)

URL: https://github.com/ComposioHQ/awesome-claude-skills

Integrates Claude with 500+ applications (e.g., Gmail, Slack, GitHub, Notion).

Categories include document processing, development tools, business automation, and creative workflows.

Each skill folder contains runnable examples and integration guides.

Antigravity Awesome Skills

URL: https://github.com/sickn33/antigravity-awesome-skills

Provides over 1,410 skills for Claude Code, Cursor, Gemini CLI, Codex, and other AI coding tools.

Organized into 9 major categories and role‑based starter bundles (e.g., Web engineer, security engineer).

Cross‑platform support and a web app for searching and filtering skills.

Agent Skills for Context Engineering

URL: https://github.com/muratcankoylan/Agent-Skills-for-Context-Engineering

Referenced by a Peking University research paper as foundational work on static skill architecture.

Covers context fundamentals (when to enter the context window and why), memory systems (short‑term, long‑term, graph‑based), multi‑agent patterns (orchestrator, peer‑to‑peer, hierarchical), and an evaluation/debugging framework.

Each skill includes Python scripts and examples for building production‑grade agents.

UI UX Pro Max Skill

URL: https://github.com/nextlevelbuilder/ui-ux-pro-max-skill

Generates a complete design system from a product description.

Inference engine contains 161 industry‑specific rules covering SaaS, beauty‑medical, fintech, etc.

Supports 15 technology stacks (React, Next.js, SwiftUI, Flutter, Tailwind, …).

Outputs include style guidelines, color palettes, typography, and accessibility checklists.

Humanizer

URL: https://github.com/blader/humanizer

Removes AI‑generated text fingerprints based on WikiProject AI Cleanup’s “AI writing signs” guide.

Detects 29 patterns (e.g., inflated importance statements, verb avoidance, excessive dashes) and provides before/after examples.

Two‑pass rewrite with style calibration to match a user’s own writing sample.

Supports Claude Code and OpenCode.

Last30Days Skill

URL: https://github.com/mvanhorn/last30days-skill

Scans Reddit and X for the past 30 days on any topic.

Returns engagement metrics (upvotes, likes, retweets) and generates ready‑to‑copy prompts.

Example: for a legal query, extracts the main theme, proposes five prompt strategies, and outputs a structured prompt with role assignment and humility instructions.

Requires an OpenAI API key for Reddit and an xAI API key for X.

Claude Scientific Skills

URL: https://github.com/K-Dense-AI/scientific-agent-skills

Contains 140 scientific skills across 12+ disciplines (bioinformatics, genomics, cheminformatics, drug discovery, proteomics, clinical research, medical imaging, …).

Integrates 28+ databases (PubMed, ChEMBL, UniProt, ClinVar, COSMIC, DrugBank, …).

Automatically installs 55+ Python packages (e.g., RDKit, Scanpy, DeepChem) required for the workflows.

OpenAI Skills

URL: https://github.com/openai/skills

Official skill directory for the OpenAI Codex CLI, illustrating the reference implementation of skills.

Organized into three tiers:

.system – auto‑installed shell/file tools.

.curated – vetted by OpenAI.

.experimental – community contributions under testing.

Includes a built‑in $skill-installer for discovery and one‑click setup.

Supports multiple languages (Python, JavaScript, Shell, Swift, PowerShell) and follows the agentskills.io open standard.