zipbox-browser

Automates headless Chromium browser sessions via Playwright CLI for JavaScript-rendered page interaction.

1|Updated Jun 17, 2026
One-click install
npx skills add https://github.com/tribes-protocol/trading-harness --skill zipbox-browser-tribes-protocol
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: zipbox-browser
Source: https://github.com/tribes-protocol/trading-harness/tree/main/skills/zipbox-browser
Command: npx skills add https://github.com/tribes-protocol/trading-harness --skill zipbox-browser-tribes-protocol

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Many websites block plain HTTP extraction with JavaScript rendering requirements, bot interstitials, or 401/403/429 responses, and some tasks demand real page interaction like clicking, typing, screenshots, or PDF capture. This Skill drives the sandbox's baked Chromium through Microsoft's Playwright CLI so those pages become accessible and interactive. ## Core Features & Use Cases - Full page interaction: Navigate, snapshot the DOM, click, fill forms, press keys, manage tabs, and handle native dialogs in named sessions. - Artifact capture and inspection: Save screenshots and PDFs, read console messages, inspect network requests, and mock routes for testing. - Desktop browser attachment: Attach over CDP to the user's visible desktop Chromium so logins and actions happen in a browser the user can watch. - Use Case: A page returns a bot-check interstitial to plain HTTP fetches. Open it in a named Playwright session, wait for a stable selector, snapshot the DOM, click through the consent banner, and extract the rendered text. ## Quick Start Use the zipbox-browser skill to open the target URL in a Playwright session, wait for the main content to render, and extract the page text.

Frequently Asked Questions about zipbox-browser

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I scrape a JavaScript-rendered page with Playwright CLI?▼

Open the URL in a named session with playwright-cli, wait for a stable selector using run-code, then capture a snapshot or evaluate document.body.innerText to extract the rendered content. Close the session when finished.

When should I use a headless browser instead of HTTP fetching?▼

Use a browser when HTTP extraction returns 401, 403, 406, or 429 errors, a bot or CDN interstitial, or empty skeleton HTML, or when the task requires clicking, typing, screenshots, or PDF capture. For plain readable pages, simple web search extraction is sufficient.

How do I attach Playwright to an existing Chrome browser over CDP?▼

Start the desktop service with zipbox-desktop up, then run playwright-cli with the attach command pointing at the CDP endpoint on http://127.0.0.1:9222. All subsequent commands in that session operate on the visible desktop browser.

Why does Playwright fail to launch Chromium as root?▼

Root launches fail without the sandbox disabled and an explicit executable path. Define a wrapper that sets PLAYWRIGHT_MCP_EXECUTABLE_PATH to /usr/bin/chromium and PLAYWRIGHT_MCP_SANDBOX to false so Chromium starts correctly.

What should I do when a Playwright command reports a modal state error?▼

A native dialog such as alert or confirm is pending and blocks all other commands. Run dialog-accept or dialog-dismiss to answer it, then retry the original command; the session resumes normally.

Can Playwright CLI save and reuse login sessions?▼

Yes, state-save writes cookies and storage to a JSON file and state-load restores them in another session, so authenticated pages can be revisited without signing in again. Never print the state file contents since they contain live tokens.