Skills for your coding agent
One skill per practice, for Claude Code, Codex, Cursor, Copilot or Gemini CLI, so your agent knows what good looks like while it builds. Each says how to do it and what passes our check, and links a real file from the best on our boards to read first. They're generated from our best practices, so the two always say the same thing.
Each line adds the skill for Claude Code. For the other agents, save the same file under .agents/skills/ instead. On Windows PowerShell, type curl.exe instead of curl. Your agent is told to treat the linked files as examples, not instructions, and to write only what's true of your product.
llms.txt
Write an llms.txt that lists your docs, keep its links working, and point to it from your pages.
Covers 3 checks: llms.txt, llms.txt links work, and it fits, llms.txt linked from your pages.
curl --create-dirs -fsSL https://leaderfile.vercel.app/skills/legible-llms-txt/SKILL.md -o .claude/skills/legible-llms-txt/SKILL.mdLet agents in and find you
robots.txt that lets AI agents in, a sitemap, pages readable without JavaScript, real 404s, a meta description, structured data and readable pricing.
Covers 9 checks: AI agents allowed, homepage reachable, Our scanner gets in, Sitemap, Readable without JavaScript, Real 404s, Meta description, Structured data, Readable pricing, Free tier in emails.
curl --create-dirs -fsSL https://leaderfile.vercel.app/skills/legible-discovery/SKILL.md -o .claude/skills/legible-discovery/SKILL.mdDocs agents can read
A docs page linked from your homepage, readable without a login, in Markdown, structured, and with code samples that name their language.
Covers 6 checks: Docs page, Docs readable without a login, Docs pages fit an agent's context, Structured docs page, Code samples say their language, Docs in Markdown.
curl --create-dirs -fsSL https://leaderfile.vercel.app/skills/legible-docs/SKILL.md -o .claude/skills/legible-docs/SKILL.mdAn API spec agents can call
A linked OpenAPI spec with an operationId on every operation and a schema for errors, auth an agent can follow, and a machine-readable discovery file.
Covers 6 checks: OpenAPI spec, Every operation has an operationId, Errors have a schema, Auth an agent can follow (proxy), Machine-readable discovery file, Send operation in the spec.
curl --create-dirs -fsSL https://leaderfile.vercel.app/skills/legible-api-spec/SKILL.md -o .claude/skills/legible-api-spec/SKILL.mdMCP server, sign-in and agent tools
An MCP server or CLI your docs name, sign-in an MCP client can use, a published agent skill, and tools for browser agents.
Covers 5 checks: MCP server, MCP server or CLI, Sign-in an MCP client can use, An agent skill is published, Tools for browser agents.
curl --create-dirs -fsSL https://leaderfile.vercel.app/skills/legible-mcp/SKILL.md -o .claude/skills/legible-mcp/SKILL.mdPackages and install lines
npm and PyPI packages shown in your docs, with an install line an agent can copy.
Covers 2 checks: npm and PyPI packages, Install shown for four languages.
curl --create-dirs -fsSL https://leaderfile.vercel.app/skills/legible-packages/SKILL.md -o .claude/skills/legible-packages/SKILL.md
To find the gaps first
The main Legible skill scans your live site, shows what fails and helps write your own version of each file. Add it with:
npx legible-cli skillEvery skill as a list for scripts: /skills/index.json. All the ways to use Legible from your terminal, editor and CI: Docs.