skills.rest
eze-is
eze-isCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

web-access

Automates web search, page fetching, and CDP browser control for AI agents.

9.0k|639|Updated Mar 18, 2026
Search Tags:#search#web scraping#chrome devtools#cdp#web automation#browser control#agent skill
One-click install
npx skills add https://github.com/eze-is/web-access --skill web-access-eze-is
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: web-access
Source: https://github.com/eze-is/web-access
Command: npx skills add https://github.com/eze-is/web-access --skill web-access-eze-is

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? AI agents' built-in web tools (WebSearch, WebFetch) lack scheduling strategy and browser automation, so they fail on login-required sites, anti-scraping platforms like Xiaohongshu or WeChat, and dynamic pages. This Skill adds a tool-selection strategy, CDP browser automation against the user's own Chrome or Edge, and per-site experience accumulation. ## Core Features & Use Cases - Automatic tool selection: Chooses among WebSearch, WebFetch, curl, Jina, and CDP browser control based on the task, with guidance on when static fetching fails and a real browser is required. - CDP Proxy browser automation: Connects to the user's everyday Chrome/Edge via remote debugging, inheriting login sessions, with endpoints for tabs, JS evaluation, clicking, file upload, scrolling, and screenshots including video frame capture. - Local bookmark/history search: The find-url.mjs script locates URLs from Chrome/Edge bookmarks and history by keyword, time window, or visit frequency, covering internal systems that public search cannot reach. - Use Case: Ask the agent to research five products across their official sites in parallel; it spawns sub-agents sharing one CDP proxy with tab-level isolation, then returns a comparison summary. ## Quick Start Ask your agent to install the skill from https://github.com/eze-is/web-access and then request any web task such as searching a topic or reading a page behind a login.

Frequently Asked Questions about web-access

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

FAQPage Schema
How do I give an AI agent full web browsing capability?▼

Install the web-access skill via npx skills add eze-is/web-access or by cloning it into your agent's skills directory. It handles search, fetching, and browser automation automatically once loaded, working with Claude Code, Cursor, Gemini CLI, and Codex CLI.

How to automate Chrome or Edge with CDP without losing login sessions?▼

Enable remote debugging in chrome://inspect or edge://inspect, then the skill's CDP proxy connects directly to your everyday browser over WebSocket. Because it uses your real browser profile, existing login sessions carry over without separate authentication.

Does web-access support Microsoft Edge or only Chrome?▼

It supports Chrome, Edge, Chromium, and Chrome Canary through the same auto-discovery mechanism. Set a default in config.env via WEB_ACCESS_BROWSER, or override per run with the --browser flag.

Why does fetching Xiaohongshu or WeChat pages fail with normal HTTP tools?▼

These platforms block static fetching even for public content, so curl and WebFetch return errors or empty pages. The skill routes such tasks to CDP browser mode, which renders pages like a real user and bypasses the static-layer restrictions.

What are the requirements and limitations of the CDP browser mode?▼

CDP mode requires Node.js 22 or later and a browser with remote debugging enabled. Automating social platforms carries account restriction or ban risk, so the documentation recommends using a secondary account for operations like posting.

Related Skills

Explore More Software Engineering Skills→
legoutlegout
3

chrome-cdp

Control local Chrome tabs via the Chrome DevTools Protocol for UI automation and data extraction.

Community
Intermediate
johnqtcgjohnqtcg
29

deep-research

Verify claims and synthesize findings with citations from web and codebase sources.

Community
Advanced
northguildnorthguild
6

research

Investigates questions against primary sources and saves cited findings as Markdown files.

Official
Basic
PAISHU-AIPAISHU-AI

playwright

Automate web browser interactions from the terminal using Playwright.

Community
Advanced