Agent skills for AI coding agents across 20 domains of human knowledge
A collection of agent skills for AI coding agents spanning 20 domains, from programming and business to artificial intelligence, security, and data engineering. Each skill is a self-contained package of instructions, reference documentation, and Python validation scripts. Skills follow the Agent Skills format and are installable via npx skills add.
Our Skills Index lists the complete catalog with domain, category, and description for every skill.
44 skills across 8 active domains: programming, business, artificial intelligence, devops, web, security, data, and design. 12 additional domains are reserved for future expansion.
See the full catalog: Skills Index
skills.sh handles registry and distribution. Python 3 runs validation via tools/validate-all.py. Node.js powers installation via npx skills add. Husky 9 and Commitlint 20 enforce conventional commits on every change. Bun is used for local development.
Eight domains are active: 01-programming (21 skills), 02-business (1), 03-artificial (5), 04-devops (1), 05-web (7), 06-security (4), 07-data (1), and 08-design (4). Domains 09 through 20 are reserved for future expansion.
Every push and pull request is automatically validated via GitHub Actions to ensure all skills remain compliant.
Prerequisites: Node.js 18+, Python 3.8+, and optionally Bun.
# Install all skills
npx skills add natuleadan/skills
# Install a specific skill
npx skills add natuleadan/skills --skill 010101-package-securityFor local development: git clone, bun install, then python3 tools/validate-all.py.
Skills use a flat directory structure with six-digit codes that encode a three-level hierarchy: domain (first 2 digits), category (next 2), and skill (last 2). The naming format is DDCCSS-word-word — exactly two hyphenated words after the code, each at least four characters. Every skill lives directly under skills/ with a SKILL.md, scripts/, and optional references/ directory.
The manifest at .claude-plugin/marketplace.json is the single source of truth for what the repo contains. Our Contribution Guidelines document provides full conventions for skill creation.
Commit messages use conventional format: type(scope): description. Supported types include feat, fix, upgrade (breaking), docs, refactor, test, chore, ci, and others. Scope is mandatory, subject is lowercase, maximum 100 characters.
Skill names: lowercase, hyphens only, maximum 64 characters, must match the directory name. All content is written in English. Frontmatter requires name and description; optional fields include license, compatibility, metadata, and allowed-tools.
Every skill is validated through python3 tools/validate-all.py which checks registry integrity, frontmatter compliance, directory-to-name matching, metadata.json completeness, and empty files. CI runs this on every push. Our validation process is documented in CONTRIBUTING.md.
Plan the skill domain and functionality, create the directory with SKILL.md + scripts/ + references/, register the path in marketplace.json, run validate-all.py, commit with a conventional message, and push. Husky runs validation automatically on every commit.
Our Contribution Guidelines document provides detailed conventions for skill creation, covering directory structure, SKILL.md frontmatter rules, naming, validation, and commit format. All contributions must pass python3 tools/validate-all.py before merging.
MIT — Leonardo Jara. See LICENSE for full terms.
Maintained by Leonardo Jara (@leojara95) under the natuleadan organization (@natuleadan).
