webforJ AI Plugin
The webforJ AI plugin is the recommended way to connect your AI coding assistant to webforJ. One install gives your assistant the full toolkit: live access to webforJ docs, project scaffolding, theme generation, design-token validation, and structured workflows that teach it how to use all of it correctly.
What you get
Installing the plugin connects two complementary pieces in a single step:
- webforJ MCP server - live tools the assistant can call on demand: look things up in the webforJ knowledge base, scaffold Maven projects, generate DWC themes, read the styling surface of any DWC component, and validate
--dwc-*tokens before they land in your CSS. - Agent Skills - structured workflows that tell the assistant when to reach for those tools, what order to do things in, and how to validate the result. Covers building reusable components and styling webforJ apps end-to-end.
Together they turn an AI assistant that guesses at webforJ conventions into one that follows them.
Even with the plugin, AI assistants can produce incorrect code in complex scenarios. Always review and test generated code before shipping.
Installation
- Claude Code
- GitHub Copilot CLI
- VS Code + Copilot
- Gemini CLI
- OpenAI Codex CLI
claude plugin marketplace add webforj/webforj-ai
claude plugin install webforj@webforj-ai
Verify inside Claude Code:
/plugin
/mcp
The webforj plugin appears under Installed. The MCP server appears as plugin:webforj:webforj-mcp under connected servers.
copilot plugin marketplace add webforj/webforj-ai
copilot plugin install webforj@webforj-ai
Verify:
copilot plugin list
From the command palette, run Chat: Install Plugin From Source, then paste:
webforj/webforj-ai
gemini extensions install https://github.com/webforj/webforj-ai
Verify:
gemini extensions list
codex plugin marketplace add webforj/webforj-ai
Then open a Codex session, run /plugins, select webforj, and press Space to enable it.
Codex doesn't auto-load skills by prompt match like other clients. Invoke them explicitly: Codex doesn't auto-load skills by prompt match like other clients. Invoke them explicitly:
$webforj:webforj-styling-apps theme this app with a blue palette
$webforj:webforj-creating-components wrap this Custom Element as a webforJ component
MCP tools work automatically without the $ prefix.
Other clients
Cursor, Kiro, Goose, Junie, Antigravity, and any other Agent Skills compatible client support the plugin too - they just use manual configuration rather than a marketplace command. See the per-client install guide for the exact steps.
Using it
Once installed, most assistants load the right piece automatically based on your prompt:
- "Wrap this Custom Element library as a webforJ component." - triggers the creating-components skill
- "Style this view with the DWC design tokens." - triggers the styling-apps skill
- "Scaffold a new webforJ sidemenu project called CustomerPortal." - calls the MCP project scaffolder
- "Generate a theme from brand color
#6366f1." - calls the MCP theme generator - "Find the webforJ docs on
@Routeand routing." - calls the MCP knowledge search
For best results, always mention webforJ in your prompts - that's the cue the assistant uses to reach for the plugin instead of general Java knowledge.
Updating and uninstalling
Each supported client has its own update and uninstall commands. See the webforj-ai README for per-client instructions.