
WebMCP Tool Generator
Builds and tests experimental WebMCP tools that let agents use a web page's real workflows safely.
- Repository
- vercel-labs/agent-browser
- Stars
- 45,202
- Licence
- Apache-2.0
- Files
- 1 file
What it does
This skill teaches an agent to turn an existing web page's real workflows into WebMCP tools, a proposed way for pages to expose actions to AI agents. The agent explores the page with the agent-browser tool, writes a manifest, a small init script (webmcp.init.js) and evaluation cases, then checks that each tool works and produces the same visible result as the normal UI. It also compares the tool against the standard accessibility-tree approach where an external agent runtime is available, and records the results in a report.
Great for
- Web developers experimenting with WebMCP on forms or workflows they already own
- Teams comparing WebMCP tools against accessibility-tree browser automation
- QA engineers who want repeatable deterministic checks for agent-facing page actions
Skip it if
- Production sites that need stable, supported agent integrations, because WebMCP is experimental and the skill itself labels it so
- Pages with consequential actions such as payments or account changes unless a developer adds explicit confirmation and an independent result check, since the skill requires these but cannot guarantee them
- Non-developers with no access to a browser automation setup, since it depends on running agent-browser commands
Try asking
- “Create WebMCP tools for the search and filter workflow on my store's product listing page and save them under artifacts/shop.example/search”
- “Write webmcp.init.js for our contact form and check that invoking the tool fills and submits it the same way the UI does”
- “Compare an accessibility-tree automation of our signup flow against a WebMCP version and record success, latency and token use”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.
webmcp-gen.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip webmcp-gen.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip webmcp-gen.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Start with a read-only workflow so you can verify registration and invocation before attempting anything that changes data.
- Use declarative WebMCP for semantic HTML forms and reserve imperative tools for workflows that cannot be expressed declaratively.
- Keep fixtures free of real credentials and cookies, and review generated webmcp.init.js before loading it into a browser session.
What's inside
- SKILL.md
Redistributed unmodified from vercel-labs/agent-browser under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.