ak-agent-browser

name	ak:agent-browser
description	Browser and desktop automation through the agent-browser CLI. Use for long autonomous browsing, compact page snapshots, screenshots, form filling, login/test flows without real Chrome profile state, scraping, exploratory QA, cloud browsers, Electron apps, Slack automation, and bug hunts.
user-invocable	true
when_to_use	Invoke for browser/app automation that needs snapshots or clicks and does not require the user's real Chrome profile state.
category	dev-tools
keywords	["browser","automation","playwright","testing","e2e","browserbase","autonomous","headless","electron","slack","dogfood","agentcore","vercel-sandbox"]
license	Apache-2.0
allowed-tools	Bash(agent-browser:*), Bash(npx agent-browser:*)
argument-hint	[url or task]
metadata	{"author":"agentkit","version":"2.0.1","upstream":"vercel-labs/agent-browser"}

agent-browser Skill

Fast browser automation CLI for AI agents. Chrome/Chromium via CDP with accessibility-tree snapshots and compact @eN element refs (~280 chars/snapshot vs 8K+ for Playwright MCP).

Use ak:agent-browser for browser testing, screenshots, form fills, scraping, exploratory QA, bug hunts, cloud browsers, Electron apps, Slack automation, and flows where a fresh or tool-managed browser is fine. Prefer it over generic browser tools for profile-independent browser work.

Use ak:chrome-profile instead when the task needs the user's actual Chrome profile: existing cookies, logged-in sessions, a specific Google account, a tenant/workspace already open in daily Chrome, or deterministic targeting across multiple Chrome profiles.

Install / Upgrade

npm install -g agent-browser     # install (or upgrade) to latest
agent-browser install            # download Chromium (one-time)
agent-browser install --with-deps  # Linux: include system deps
agent-browser upgrade            # self-upgrade the binary
agent-browser --version          # verify

Re-run npm install -g agent-browser (or agent-browser upgrade) periodically — new commands and skills ship with the binary.

Start here — load live workflow content

This file is a discovery stub, not the usage guide. Before running any agent-browser command, load workflow content from the installed CLI:

agent-browser skills get core             # start here: workflows, common patterns, troubleshooting
agent-browser skills get core --full      # include full command reference and templates

The CLI serves skill content from the installed binary, so instructions always match the installed version. This stub stays intentionally small and points at skills get core instead of duplicating command details that can change between releases.

Specialized skills

Load a specialized skill when the task falls outside browser web pages:

agent-browser skills get electron          # Electron apps (VS Code, Slack, Discord, Figma, ...)
agent-browser skills get slack             # Slack workspace automation
agent-browser skills get dogfood           # Exploratory testing / QA / bug hunts
agent-browser skills get vercel-sandbox    
agent-browser skills get agentcore