Public Browser — Testing skill for Claude Code
Token-efficient Chrome browser automation MCP server for AI agents.
How to install Public Browser
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open Silbercue/public-browser and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Public Browser does
Token-efficient Chrome browser automation MCP server for AI agents. Direct CDP, a11y-tree refs, multi-tab, server-side plan executor. 5.4x smaller page snapshots than Playwright MCP. MIT, no paid tier.
Alternatives in Testing
- Design Compare — Capture two websites with Playwright and produce a structured side-by-side design comparison — layout, color 309 ★
- AI Agents Public — Production-grade agent skills and Custom GPT prompts for ChatGPT, Claude Code, and Codex 82 ★
- Monobrowse — Launch monomind browse for web UI testing and automation — native TypeScript CDP client, no external binary ne 16 ★
README
Public Browser
[](https://github.com/Silbercue/public-browser/releases) [](https://www.npmjs.com/package/public-browser) [](https://github.com/Silbercue/public-browser#tool-overview) [](LICENSE) [](https://nodejs.org)
A Chrome MCP server that talks straight to CDP. It finished the same 30-test benchmark page in 84 and 86 tool calls where Playwright MCP needed 137 and 151 — two runs each, 2026-09-03, driver Claude Opus 5, same pass rate ([Benchmarks](#benchmarks), including where it loses). Direct CDP, a11y-tree refs, multi-tab ready — 2,322 TypeScript tests, 237 Python tests.
Built for [Claude Code](https://claude.ai/claude-code), [Cursor](https://cursor.sh), and any MCP-compatible client.
**Looking for an alternative to Playwright MCP, Browser MCP, or claude-in-chrome?** Public Browser talks to Chrome directly via the DevTools Protocol — no Playwright dependency, no Chrome extension bridge, no single-tab limit. One command to install, zero config. [See benchmark comparison below](#benchmarks).
Why Public Browser?
Every Chrome MCP server has the same problem: **bulky responses, too few reliable refs.** Screenshots return 10-30x more context payload than text trees. Selector-based refs break the second the DOM rerenders. Extension bridges (Browser MCP) get stuck on the connected tab. Playwright wrappers spin up a new browser instance for every session.
Public Browser fixes this. It talks directly to Chrome via CDP (same protocol Playwright and Puppeteer use internally), returns an accessibility-tr
Related Skills
Smart Browser
Browser automation and API testing MCP server on Playwright + Chrome DevTools Protocol. Numbered element refs
Webtest Orch
Token-efficient e2e orchestration skill for Claude Code: explore once, replay deterministically. Playwright +
Marketing Design
Direct invocation of Marketing Designer. Author or revise a marketing-design-spec for a project's public marke
Handoff Package
Package an internal agent into a public-safe, reproducible handoff bundle — cleaned source, docs, examples, MI
Skill QA
Q.A./Quality Assurance da casa — a disciplina de teste agnóstica de linguagem (pirâmide, comportamento, smoke
Claude Md Taskview
A Claude Code mod that draws the milestone and task tree of your spec//tasks.md files in a pane beside the tra
Related Agents
Cdp Engineer
Persona: CDP Engineer (Chrome DevTools Protocol)
Visual Tester
Visual QA tester — navigates web UIs via Chrome CDP, spots visual issues, tests interactions, produces structu
Web Debugger
Interactive headful Chrome debugging via Chrome DevTools MCP — live DOM snapshots, console, network, and perfo