browser · stellarlinkco — Claude Code skill: what it does, how to install
browser automates Chrome actions via Chrome DevTools Protocol without configuring an MCP server. Useful for taking screenshots, running scripts, and inspecting DOM nodes.
AGPL-3.0 202 installs ★ 2,756 updated 2026-05-04
npx skills add stellarlinkco/myclaude --skill browser -a claude-codeWhat browser does
The skill equips your coding agent with lightweight Node.js utilities that communicate with Google Chrome through CDP. Everything connects directly to localhost:9222, bypassing the overhead of standalone MCP infrastructure.
Included scripts handle starting the browser, navigating URLs, evaluating JavaScript, and saving PNG screenshots into a temporary folder. The visual element finder scripts/pick.js scans page elements against a text prompt and returns structured JSON with tags, classes, selectors, and bounding boxes. Running scripts/start.js with the --profile flag retains local cookies and session tokens.
The tool fails unless you install the ws package in the skill directory before the first run. All automation stays on your machine, so credentials and session data never leave your local environment.
What it helps with
- Start Chrome with remote debugging enabled on port 9222
- Take PNG screenshots of web pages directly to a temporary directory
- Execute custom JavaScript inside an active browser tab
- Retrieve DOM element selectors and coordinates from visual descriptions
Example prompts for the agent
- Launch Chrome with persistent profile and navigate to https://example.com
- Evaluate
document.titleon the active page and display the output - Locate the form submit button and inspect its bounding box coordinates
How to install browser
1Skills CLI (Vercel)
npx skills add stellarlinkco/myclaude --skill browser -a claude-code2Manually: put the files in place
~/.claude/skills/browser/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/browser/
1Skills CLI (Vercel)
npx skills add stellarlinkco/myclaude --skill browser -a codex2Manually: put the files in place
.agents/skills/browser/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add stellarlinkco/myclaude --skill browser -a cursor2Manually: put the files in place
.cursor/skills/browser/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install stellarlinkco/myclaude skills/browser2Manually: put the files in place
.github/skills/browser/SKILL.md1Skills CLI (Vercel)
npx skills add stellarlinkco/myclaude --skill browser -a gemini-cli2Platform CLI
gemini skills install https://github.com/stellarlinkco/myclaude --path skills/browserManually: put the files in place
.devin/skills/browser/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add stellarlinkco/myclaude --skill browser -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 browser
How to install browser in Claude Code?
Run npx skills add stellarlinkco/myclaude --skill browser -a claude-code in your terminal. Next, install dependencies using npm install --prefix ~/.claude/skills/browser/browser ws.
What does the browser skill do?
It connects to Chrome over Chrome DevTools Protocol to automate browsing workflows. The agent can open links, grab page screenshots, inspect DOM elements, and run JavaScript snippets.
Does browser require an MCP server?
No, it works without an MCP server. The tool relies on direct Node.js helper scripts communicating over WebSockets with Chrome on localhost:9222.
Similar skills
All skills for Browser and scraping →browser by other authors comes first.