Vercel Labs Agent Browser
vercel-labs-agent-browser is a Claude Code plugin for the Ship phase that provides a browser automation CLI so agents can test flows, fill forms, capture screenshots, and extract page data.
by vercel-labs · github.com/vercel-labs/agent-browser
Add agent-browser when your Claude agent must drive a real browser for E2E checks, form filling, screenshots, and extraction during validate, build, and ship.
Add it to Claude Code
Install the plugin in Claude Code. One command, paste-ready.
/plugin install vercel-labs-agent-browser@vercel-labs/agent-browserBuilt to be called by your agent
Skillselion is itself an MCP server. Your agent can pull this entry and a paste-ready install config straight from the API - no copy-paste.
Retrieve this entry with skillselion.get_details("plugin:vercel-labs/agent-browser") and the paste-ready config with skillselion.get_install_config("plugin:vercel-labs/agent-browser").
What it does
vercel-labs-agent-browser is a Claude Code plugin that exposes a browser automation CLI tuned for AI agents, not for human-only test runners. Solo builders install it when the coding agent must open pages, click through flows, fill forms, grab DOM or visual evidence, and support lightweight end-to-end validation without leaving the agent session. The catalog positions it squarely in development automation: keywords highlight agent, browser, extraction, filling, interactions, screenshots, and testing. That makes it valuable during validate-phase prototypes when you want the agent to demo a landing or wizard, during build when UI integrations need live verification, and during ship when regression checks and visual snapshots gate release. It is narrower than orchestration platforms—one tool, one job—so setup stays closer to beginner-intermediate if you already run Claude Code plugins and a controllable browser environment. Pair it with review and security practices when automating authenticated flows so credentials and production data stay scoped.
Highlights
- Browser automation CLI purpose-built for AI agents (Claude Code plugin packaging)
- Automates real browser interactions including navigation, filling, and structured extraction
- Supports screenshot capture for visual verification and agent-grounded debugging
- Fits agent-driven testing loops without maintaining a separate Playwright script repo for every task
- Single focused plugin (1 entry) from vercel-labs rather than a broad multi-agent catalog
Why builders use it
Agents can write frontend code but often cannot verify it in a real browser without brittle one-off scripts or manual QA.
After you add the plugin, your agent can run browser automation from the CLI for testing, extraction, and visual checks aligned with ship and validate workflows.
At a glance
- Type - Plugin in Automation.
- Adoption - 0 installs, 27.6k stars, 0 votes.
FAQ
Who is vercel-labs-agent-browser for?
It is for AI-coding workflows on Claude Code where the agent must control a browser for testing, forms, screenshots, and data extraction.
When should I use vercel-labs-agent-browser?
Use it when validating UI prototypes, running agent-led E2E checks, or capturing page evidence before launch without hand-writing a new automation script each time.
How do I add vercel-labs-agent-browser to my agent?
Install the vercel-labs/agent-browser Claude Code plugin from the catalog and invoke the browser CLI through your agent’s plugin command surface per project docs.
Comments
Share how you use vercel-labs-agent-browser, gotchas, or tips for other indie builders.
No comments yet - be the first to share how you use it.