browse · ulpi-io — Claude Code skill: what it does, how to install
Controls a persistent Chromium instance and native apps via the browse CLI. Useful for testing UI workflows, verifying forms, and inspecting live pages.
no license 282 installs ★ 5 updated 2026-08-31
npx skills add ulpi-io/skills --skill browse -a claude-codeWhat browse does
The skill instructs an active browse CLI instance directly from the terminal. Rather than fetching static HTML, it maintains a persistent Chromium process that retains cookies and local state across consecutive actions. Beyond standard web flows and local dev servers, it supports Camoufox for protected targets alongside macOS, iOS, and Android interfaces.
Interaction relies on @eN node references captured with snapshot -i rather than brittle CSS selectors. When the DOM changes, a fresh snapshot gives new references so the agent does not click stale coordinates.
The browse binary must exist on your PATH before running commands. Native app targets do not support goto, eval, or cookie management; those targets work strictly through accessibility tree events.
What it helps with
- test login flows on a local development server
- capture interactive DOM element snapshots using @eN references
- automate mobile UI testing in iOS Simulator or Android Emulator
- inspect runtime console errors and network traffic in a live browser
Example prompts for the agent
- Navigate to
http://localhost:3000and inspect network errors on submit - Take an interactive snapshot with
browse snapshot -iand click ref @e2 - Start the Android emulator with package
com.android.settingsand check the screen
How to install browse
1Skills CLI (Vercel)
npx skills add ulpi-io/skills --skill browse -a claude-code2Manually: put the files in place
~/.claude/skills/browse/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/browse/
1Skills CLI (Vercel)
npx skills add ulpi-io/skills --skill browse -a codex2Manually: put the files in place
.agents/skills/browse/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add ulpi-io/skills --skill browse -a cursor2Manually: put the files in place
.cursor/skills/browse/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install ulpi-io/skills browse2Manually: put the files in place
.github/skills/browse/SKILL.md1Skills CLI (Vercel)
npx skills add ulpi-io/skills --skill browse -a gemini-cli2Platform CLI
gemini skills install https://github.com/ulpi-io/skills --path browseManually: put the files in place
.devin/skills/browse/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add ulpi-io/skills --skill browse -a opencodeClaude.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 browse
How to install browse in Claude Code?
Run npx skills add ulpi-io/skills --skill browse -a claude-code. The skill file installs to ~/.claude/skills/browse/SKILL.md.
What does browse do?
It allows the agent to control a live Chromium browser and simulator apps using the browse CLI tool. The agent can submit forms, navigate pages, read logs, and capture UI snapshots.
Can browse evaluate JavaScript inside iOS or Android apps?
No, commands like js, goto, eval, and cookies are restricted to browser targets. Native mobile and desktop apps only handle accessibility actions like tap, fill, and swipe.
Similar skills
All skills for Browser and scraping →browse by other authors comes first.