| name | use-browser |
| description | Browser automation using Chrome DevTools MCP and the agent-browser CLI. Use when the user needs to navigate websites, fill forms, take screenshots, scrape data, or test web applications. |
| allowed-tools | Bash(agent-browser:*) |
Browser Use
Two complementary approaches:
- Chrome DevTools MCP — tools prefixed
mcp__chrome-devtools__* against a running Chrome instance (when available)
- agent-browser CLI — standalone headless browser for CI, cloud, recording, and environments without local Chrome
When to Use
- Navigate sites, fill forms, take screenshots, scrape or inspect pages
- Record demos, compare pages, run Lighthouse / performance traces
- QA scanning and product research that need a real browser
Tool Selection
| Prefer Chrome DevTools MCP when | Prefer agent-browser CLI when |
|---|
| Live inspection of the user’s Chrome session / auth | Video recording (record start/stop) |
| Snapshots, screenshots, JS eval, console/network debug | Session/profile persistence (--session, --profile) |
| Form interaction + device emulation | Headless-only / cloud (Browserbase, Browserless, …) |
| Lighthouse and performance traces | PDF export, annotated screenshots, snapshot/screenshot diffs |
No universal best tool — pick by Chrome availability, recording needs, and environment constraints.
Connection Setup
MCP → running Chrome
Preferred (autoConnect): open Chrome → chrome://inspect#remote-debugging → enable remote debugging. MCP connects with no extra config.
Manual: start Chrome with --remote-debugging-port=9222, then point the MCP server at --browser-url=http://127.0.0.1:9222.
/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome --remote-debugging-port=9222 &
Standalone (agent-browser only)
agent-browser open <url>
agent-browser open <url> --ignore-https-errors
agent-browser open <url> --idle-timeout 5m
Isolated headless browser; also supports Brave auto-discovery alongside Chrome/Chromium.
Core loops
Chrome DevTools MCP
- Ensure Chrome + MCP connected (above)
navigate_page / new_page → wait_for as neededtake_snapshot (or screenshot) before actingclick / fill / fill_form / keys using selectors from the snapshot- Re-snapshot after navigation or large DOM changes
- For tool names and flags → load
references/chrome-devtools-mcp.md
agent-browser CLI
agent-browser open <url>agent-browser snapshot -i → use @eN refs from the tree- Interact (
click, fill, …) with those refs
- Re-snapshot after navigation or significant DOM changes
- For full flag/subcommand catalog → load
references/agent-browser-cli.md
(if mismatch with install, prefer agent-browser --help)
Load for depth
| Need | Open |
|---|
| Full MCP tool tables + server flags | references/chrome-devtools-mcp.md |
| Full agent-browser CLI catalog | references/agent-browser-cli.md |
| Worked examples (form, record, standalone) | references/workflows.md |
Related Skills
- qa — app scanning and visual inspection
- product — competitor research and screenshot comparison
External docs
