skills.rest
luojiaaoo
luojiaaooCOMMUNITY
Explore all skills by this author →
Website𝕏@luojiaaooGitHub

agent-browser

Automates browser navigation, form filling, and data extraction through CDP-connected sessions.

8|2|Updated Apr 29, 2026
Search Tags:#cli#browser automation#web testing#web scraping#form filling#cdp#screenshot
One-click install
npx skills add https://github.com/luojiaaoo/xpeech --skill agent-browser-luojiaaoo
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/luojiaaoo/xpeech/tree/main/xpeech/agent/skills/buildin/agent-browser
Command: npx skills add https://github.com/luojiaaoo/xpeech --skill agent-browser-luojiaaoo

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires agent-browser.

What problem does it solve? It lets an AI agent control a real browser to navigate websites, click elements, fill forms, extract page data, and take screenshots without manually managing browser installations or sessions. ## Core Features & Use Cases - CDP-Based Browser Control: Connects to an externally managed browser via an injected Chrome DevTools Protocol endpoint, so no local browser installation or startup is required. - Accessibility-Tree Interaction: Uses accessibility snapshots and compact @eN element references to reliably click, type, and inspect page elements. - Version-Matched Workflows: Loads usage instructions matching the installed CLI version via agent-browser skills get core. - Use Case: Ask the agent to open a web application, log in through a form, navigate to a report page, extract the table data, and capture a screenshot as evidence. ## Quick Start Use the agent-browser skill to open https://example.com, take an accessibility snapshot, and click the first link.

Frequently Asked Questions about agent-browser

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

FAQPage Schema
How do I automate browser tasks with agent-browser?▼

Run commands like agent-browser open, snapshot, click, and screenshot after loading the matching workflow with agent-browser skills get core. Session and CDP endpoint parameters are injected automatically, so you never pass them manually.

How do I click elements on a web page using a CLI?▼

Take an accessibility snapshot with agent-browser snapshot to get compact @eN element references, then run agent-browser click @e1 or the relevant reference. This avoids brittle CSS selectors and matches the page's accessibility tree.

Does agent-browser require installing Chrome or Playwright?▼

No. The browser is provided externally through an injected CDP connection, and the skill explicitly forbids installing Chrome, Chromium, Playwright, Puppeteer, or any browser binaries. You also must not start or stop local browsers.

What should I do if the CDP connection fails?▼

Report the connection error instead of falling back to a local browser. The skill's constraints prohibit launching or managing local browsers, so a failed CDP endpoint must be fixed at the infrastructure level.

When should I take screenshots during browser automation?▼

Take screenshots only when visual inspection or evidence is useful, such as verifying a rendered layout or documenting a result. For routine interactions, prefer accessibility snapshots and @eN references, which are more reliable and compact.

Related Skills

Explore More Software Engineering Skills→
9999-12-319999-12-31

browser_cdp

Connect to and manage Chrome/Chromium browsers via CDP with port scanning.

Official
Advanced
buildinternetbuildinternet
3

finding-changelogs

Evaluate changelog URLs and recommend optimal ingestion methods.

Official
Advanced
imMamdouhaboammarimMamdouhaboammar

web-to-markdown

Converts JavaScript-rendered web pages to clean Markdown using Puppeteer and Readability.

Community
Intermediate
dapperdiversdapperdivers
1

web-browser

Automate browser interactions with dynamic websites via Chrome DevTools Protocol.

Community
Intermediate