Evil Martians Agent Skills

Agent Skills we build and use at Evil Martians, packaged so you can install them into your own AI coding agent.

A skill is a folder with a SKILL.md in it: a name, a description that tells the agent when to reach for it, and the instructions to follow once it does. The format is shared across agents, so these work the same in Claude Code, Codex, Cursor, Gemini CLI, and GitHub Copilot.

Browse them at https://evilmartians.com/agent-skills.

Agent Skills are built by Evil Martians, an American design and engineering consultancy for developer tools, AI, and cybersecurity startups.

Skills

Skill	What it does
good-readme	Writes or improves a README.md for an open source project, based on Evil Martians' real experience of successfully promoting PostCSS, Nano ID, and other popular projects. Companion to How to make your open source popular.
llms-visibility	Makes a site visible and readable to LLMs and AI agents. Covers Markdown routes, Accept content negotiation, llms.txt, and crawler signals in robots.txt. Pushes back on popular patterns that don't work. Companion to How to make your website visible to LLMs.
secure-npm-package	Sets up a secure npm release process — Trusted Publishing with staged publishes from GitHub Actions, hardened workflows, dependency cooldowns — and walks the user through the manual npmjs.com and GitHub settings with links resolved for their repo. Companion to The secure way to release an npm package in 2026.
skills-visibility	Publishes a catalog of agent skills and makes it discoverable and installable by AI coding agents — a .well-known discovery index, self-hosted archives with integrity digests, and several install commands. Pushes back on pointing an index at raw GitHub URLs. Companion to Ship agent skills like packages: discovery index, digests, and install sources.
tailwind-best-practices	Writes and audits Tailwind CSS with the five practices that keep a fast-to-write codebase maintainable: design tokens over magic values, short and consistently ordered class lists, semantically grouped tokens, and fixed variants for design-system components instead of arbitrary className props. Companion to 5 best practices for preventing chaos in Tailwind CSS.

Install

Skills install one at a time. Swap llms-visibility for whichever one you want.

npx skills add evilmartians/agent-skills --skill llms-visibility -a claude-code -g

Pass your agent to -a: claude-code, codex, cursor, gemini-cli, or github-copilot. Drop -g to install into the current project instead of your home directory.

Claude Code can install a skill from this repo's plugin marketplace instead:

claude plugin marketplace add evilmartians/agent-skills
claude plugin install llms-visibility@evilmartians

Alternatively, GitHub CLI's gh skill can install a skill and prompts you to pick which agent to install it into:

gh skill install evilmartians/agent-skills llms-visibility

Or copy the skill folder into a skills directory yourself: .claude/skills for Claude Code, .codex/skills for Codex, or the agent-agnostic .agents/skills, which any agent reads. Prefix the path with ~/ to install it globally, or use it at a project root to keep it local. Copy the whole folder, not just the SKILL.md, since a skill can ship scripts, references, and other assets alongside it.

Adding a skill

We curate this repo ourselves and don't take outside contributions. The skills are MIT licensed, so fork one and make it your own if you want to.

Maintainers: add skills/<slug>/SKILL.md with name (the same as the directory) and description in the front matter, then add a matching entry to the marketplace manifest and a row to the Skills table above. Whatever else the skill needs at runtime goes in that same folder and ships with it. CI checks the front matter and the manifest, but the table is on you.

License

MIT © Evil Martians