agent-ready-webmcp — Claude Code skill: what it does, how to install
agent-ready-webmcp wires a site's actions into the WebMCP API so browser-based AI agents can find and call them.
Apache-2.0 ★ 141 updated 2026-10-29
npx skills add fabricioctelles/skills --skill agent-ready-webmcp -a claude-codeWhat agent-ready-webmcp does
The skill walks through calling document.modelContext.registerTool() for each action an agent should reach, such as search, navigation or data lookups. Every tool needs a name, description, a JSON Schema inputSchema, and an execute callback, and the registration call returns a Promise.
Existing HTML forms can skip the JS call entirely: adding toolname and tooldescription attributes to the <form> tag, plus toolparamdescription on its inputs, exposes them declaratively. An AbortController signal passed in the options lets tools be unregistered later. Older Chrome builds put the API on navigator.modelContext instead, so code should check document.modelContext first and fall back if it's missing.
Browsers detect the API only while loading the page, so the registration script has to run on page load. To check the result, send POST https://isitagentready.com/api/scan with {"url": "https://YOUR-SITE.com"} and look for checks.discovery.webMcp.status set to "pass".
What it helps with
- expose a site's actions to browser-based AI agents
- register search and navigation tools through WebMCP
- turn existing HTML forms into declarative agent tools
- unregister tools with an AbortController signal
- confirm WebMCP support with the isitagentready.com scanner
Example prompts for the agent
- Register a search tool with document.modelContext.registerTool
- Expose this form to agents using toolname and tooldescription attributes
- Run a scan to check if the site passes the WebMCP discovery check
How to install agent-ready-webmcp
1Skills CLI (Vercel)
npx skills add fabricioctelles/skills --skill agent-ready-webmcp -a claude-code2Manually: put the files in place
~/.claude/skills/webmcp/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/webmcp/
1Skills CLI (Vercel)
npx skills add fabricioctelles/skills --skill agent-ready-webmcp -a codex2Manually: put the files in place
.agents/skills/webmcp/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add fabricioctelles/skills --skill agent-ready-webmcp -a cursor2Manually: put the files in place
.cursor/skills/webmcp/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install fabricioctelles/skills skills/agent-ready-cloudflare/webmcp2Manually: put the files in place
.github/skills/webmcp/SKILL.md1Skills CLI (Vercel)
npx skills add fabricioctelles/skills --skill agent-ready-webmcp -a gemini-cli2Platform CLI
gemini skills install https://github.com/fabricioctelles/skills --path skills/agent-ready-cloudflare/webmcpManually: put the files in place
.devin/skills/webmcp/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about agent-ready-webmcp
How do I install agent-ready-webmcp in Claude Code?
Run npx skills add fabricioctelles/skills --skill webmcp -a claude-code; the skill lands in ~/.claude/skills/webmcp/SKILL.md.
What does agent-ready-webmcp do?
It guides exposing a site's key actions to AI agents through the WebMCP API, registering each tool with a schema and an execute function the agent can call.
How do I confirm WebMCP works after setting it up?
Send POST https://isitagentready.com/api/scan with {"url": "https://YOUR-SITE.com"} and check that checks.discovery.webMcp.status returns "pass".
Source text
SKILL.md · Apache-2.0
---
name: agent-ready-webmcp
description: >
Sub-skill de agent-ready-cloudflare: Implement WebMCP
---
# Implement WebMCP
Expose site tools to AI agents via the browser using the
[WebMCP API](https://webmachinelearning.github.io/webmcp/)
([Chrome blog](https://developer.chrome.com/blog/webmcp-epp),
[declarative API](https://github.com/webmachinelearning/webmcp/blob/main/declarative-api-explainer.md)).
## Requirements
- Call `document.modelContext.registerTool()` for each tool you want to expose; it returns a Promise
- Each tool needs `name`, `description`, `inputSchema` (JSON Schema), and an `execute` callback
- Tools should expose your site's key actions (search, navigation, data retrieval)
- Pass an `AbortController` signal in the options to unregister tools when no longer needed
- For existing HTML forms, add `toolname` and `tooldescription` to the `<form>` and `toolparamdescription` to its inputs to expose them declaratively
- Older Chrome builds expose the API on `navigator.modelContext`; feature-detect `document.modelContext` first and fall back to it if needed
- The API is detected by loading the page in a browser — ensure the script runs on page load
## Validate
```
POST https://isitagentready.com/api/scan
Content-Type: application/json
{"url": "https://YOUR-SITE.com"}
```
Check that `checks.discovery.webMcp.status` is `"pass"`.
© fabricioctelles · Apache-2.0 · fabricioctelles/skills