page-agent

Name: page-agent
Availability: InStock
Author: eddielueng

Control in-page web UI elements via natural-language instructions from the DOM.

1|Updated Apr 14, 2026

Search Tags:#javascript #ui automation #web #copilot #openai-compatible #dom #in-page agent

Ready to run

npx skills add https://github.com/eddielueng/hermes-agent-zh --skill page-agent-eddielueng

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: page-agent
Source: https://github.com/eddielueng/hermes-agent-zh/tree/main/optional-skills/web-development/page-agent
Command: npx skills add https://github.com/eddielueng/hermes-agent-zh --skill page-agent-eddielueng

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

page-agent solves the problem of turning time-consuming, click-by-click web UI navigation into natural-language instructions that drive the UI directly in the browser.

Core Features & Use Cases

In-page natural language UI control: Read the current page DOM as text and execute instructions like “click login, then fill username as John” without screenshots or a browser extension.
Embeddable copilot for existing web apps: Add an AI assistant to a SaaS/admin panel/B2B tool or modernize a legacy UI without rewriting the frontend.
Evaluation and accessibility scenarios: Use it to evaluate different LLM endpoints (local or cloud) and help users perform tasks through voice/screen-reader-friendly text prompts.

Quick Start

Install page-agent into your web app, then tell it in plain language what you want the user interface to do.

Frequently Asked Questions about page-agent

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

FAQPage Schema

How do I add an in-page AI copilot to control a web UI with natural language?▼

To add an in-page AI copilot for natural language UI control, you embed a client-side script that reads the current page DOM as text and executes user-like actions based on plain language instructions.

Can I automate web UI workflows without using headless browser automation?▼

Yes, you can automate web UI workflows without headless browser automation by using an embedded in-page agent that interprets the live DOM and performs actions directly within the current browser page.

Do I need an OpenAI-compatible LLM endpoint for in-page UI automation?▼

Yes, in-page UI automation requires an OpenAI-compatible LLM endpoint to process natural language instructions and map them to DOM interactions for executing user-like actions.

What is the best way to modernize a legacy web frontend without rewriting it?▼

A practical way to modernize a legacy web frontend without rewriting it is embedding a copilot script that interprets the existing DOM and guides users through multi-step UI workflows.

Does in-page UI automation work for evaluating different LLM endpoints?▼

Yes, in-page UI automation works for evaluating different LLM endpoints by routing natural language commands through the embedded client-side script to interact with the live DOM.

Can I use an in-page agent for server-side browser automation?▼

No, an in-page agent cannot be used for server-side browser automation; it runs fully client-side and is designed exclusively for UI interaction within the current webpage.