bestax-bulma with LLMs

@allxsmith/bestax-bulma ships LLM-optimized documentation so AI coding agents — Claude Code, Cursor, GitHub Copilot, ChatGPT — can read the docs in full and build with the library correctly. This page explains what's published and how to use it.

Using bestax docs with AI tools

Point your assistant at the docs — the approach is the same across Claude Code, Cursor, Copilot, and ChatGPT:

• Give it the index. Add https://bestax.io/llms.txt to your project docs / rules, or paste it into the chat, so the model can discover every page and fetch what it needs.
• Feed it everything. For a one-shot load of the whole library, use https://bestax.io/llms-full.txt.
• Fetch a single page. For a focused question, link the page's Markdown directly — e.g. https://bestax.io/docs/api/elements/button.md — to keep the context small.

Skills

Beyond the raw docs, bestax ships Agent Skills that teach an agent how to build with the library (conventions, patterns, and a component catalog). Install one with the skills CLI:

npx skills add https://github.com/allxsmith/bestax --skill bestax-custom-component
npx skills add https://github.com/allxsmith/bestax --skill bestax-form
npx skills add https://github.com/allxsmith/bestax --skill bestax-icons
npx skills add https://github.com/allxsmith/bestax --skill bestax-layout-scaffold
npx skills add https://github.com/allxsmith/bestax --skill bestax-migrate
npx skills add https://github.com/allxsmith/bestax --skill bestax-optimize
npx skills add https://github.com/allxsmith/bestax --skill bestax-theming

Starting a new app? pnpm create bestax@latest offers to preinstall these skills into the generated app's .claude/skills/ (alongside a CLAUDE.md and a .claude/launch.json that lets Claude Code's browser preview start the dev server by name), so a Claude Code session picks them up automatically. See the Skills overview for what each one does.

In the npm package

The published @allxsmith/bestax-bulma tarball also carries small pointer files at the package root, so an agent that explores node_modules by filename (find / ls for AGENTS.md, CLAUDE.md, llms.txt) lands on these resources even if it never opens the README or reaches the network first:

File	What it is
llms.txt	A stub index pointing at the site artifacts above, plus the compact security-posture summary.
AGENTS.md	The same links, the core library conventions, and the same security-posture summary.
CLAUDE.md	A copy of AGENTS.md under the filename Claude-family tooling probes for first.

They are pointers plus a compact, verifiable security-posture summary — the site artifacts stay the source of truth for everything else, so the rest of the tarball cannot go stale between releases.

MCP server

bestax-mcp is the first-party Model Context Protocol server. Where llms.txt gives an agent the docs to read, the MCP server lets it ask questions — every component's props (including compound parts like Navbar.Brand), the working examples, the --bulma-* variables behind each component, the helper props that replace inline styles, and the Agent Skills as invocable prompts.

Setup

Each release is published to the official MCP Registry as io.github.allxsmith/bestax-mcp, for clients that install from it.

Claude Code:

claude mcp add bestax -- npx -y bestax-mcp@1

Cursor, Claude Desktop, Windsurf, Cline — add to your MCP config (.cursor/mcp.json, claude_desktop_config.json, …):

{
  "mcpServers": {
    "bestax": {
      "command": "npx",
      "args": ["-y", "bestax-mcp@1"]
    }
  }
}

Run it from your project directory, so it can find your installed @allxsmith/bestax-bulma.

The @1 pins the major version. Without it, npx resolves whatever is newest on every launch — so a breaking change, or a compromised release, reaches your agent the next time it starts, with no review step. Add it to your project's dev dependencies instead if you would rather your lockfile decide.

Tools

Start with list_components — its output names the tool to call next. Reach for search_bestax when you need a component whose name you do not know.

Tool	What it gives the agent
list_components	Every component with a one-line purpose, by category
search_bestax	Components, props, examples, CSS variables and skills in one ranked list
get_component	Import, summary and props; optionally examples, CSS variables, accessibility, related
get_props	One prop table, including compound sub-paths (Navbar.Brand)
get_examples	Working tsx examples from the component's documentation page
get_css_variables	The --bulma-* custom properties a component reads, with Sass names and defaults
get_helper_props	Spacing, colour, typography, flexbox and visibility props — the alternative to hand-written styles
lookup_bulma_classes	The bestax component and props for a Bulma class string (button is-primary), one row per class
list_skills	The Agent Skills
get_skill	A skill's instructions, or one of its reference documents

It also exposes each skill as an MCP prompt (theming, form, layout-scaffold, …) and serves bestax://catalog, bestax://components/{name} and bestax://skills/{name} as resources — useful in clients where the user attaches context (@-mentioning a resource), rather than a channel the model reaches for mid-task the way it does the tools above.

Offline, and pinned to a version

The index ships inside the package — there are no network calls, so nothing rate-limits and nothing breaks when you are offline.

It also means the server documents one specific bestax-bulma release. On startup it resolves the version actually installed in your project, and if that differs by a minor or major version it appends a warning to its answers rather than confidently describing props you do not have. Set BESTAX_MCP_NO_VERSION_CHECK=1 to turn that off.

Because the index is generated from the same source as this site — TSDoc for props, the SCSS for variables, these pages for examples — it cannot drift from the documentation you are reading.

Contributing

Found the LLM docs unclear, incomplete, or wrong for your agent? Please open an issue describing what you expected and what happened — feedback on how well the docs work with AI tools is especially welcome.

How these docs are generated

The LLM docs are generated at build time by docusaurus-plugin-llms (configured in docs/docusaurus.config.js), following the llmstxt.org standard. These artifacts are produced and served from the site root:

File	What it is
/llms.txt	Curated index — a table of contents linking every doc page (per the llmstxt.org spec).
/llms-full.txt	The entire documentation concatenated into a single plain-text file.
Per-page .md	Every page is also served as clean Markdown at <page>.md, e.g. /docs/guides/intro.md.

They are regenerated on every docs build, so they always match the deployed site.