skills.rest
denni281
denni281COMMUNITY
Explore all skills by this author →
GitHub

omnibot

Controls a real Chromium browser through the omnibot CLI for reading, clicking, filling, and verifying pages.

25|Updated Jun 8, 2026
Search Tags:#cli#browser automation#ai agent#web scraping#form filling#cdp#chromium
One-click install
npx skills add https://github.com/denni281/omnibot --skill omnibot-denni281
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: omnibot
Source: https://github.com/denni281/omnibot/tree/main/src/omnibot/skills/omnibot
Command: npx skills add https://github.com/denni281/omnibot --skill omnibot-denni281

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? AI agents cannot see or interact with live web pages that depend on login state, cookies, client-side rendering, or visible user tabs. This Skill connects agents to a real Chromium browser through the local omnibot daemon and CLI so they can read rendered pages, click controls, fill forms, navigate, and collect verifiable evidence. ## Core Features & Use Cases - Native browser operations: Read page content, snapshot interactive elements, click, fill, select, scroll, navigate, and wait using semantic locators and tab-scoped refs instead of fragile JavaScript shortcuts. - Structured fallback tiers: Escalate from semantic find to snapshot refs, selectors, DOM, mouse, JavaScript, and raw CDP only with documented evidence, keeping automation auditable. - Debug evidence collection: Capture screenshots, console logs, network requests, traces, and recordings to prove browser state or explain failures. - Use Case: An agent needs to log into a web app, fill a checkout form, and confirm the order preview. It creates a tab, runs observe-act-verify cycles per control, and stops before the final payment button unless the user explicitly confirms. ## Quick Start Ask the agent to open a website in a new browser tab with omnibot, read the page content, and verify the final URL before reporting the result.

Frequently Asked Questions about omnibot

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

FAQPage Schema
How do I let an AI agent control a real Chrome browser?▼

Install the omnibot CLI and connect the omnibot Chromium extension, then issue commands like snapshot, click, fill, and read through the local daemon. Every page-state command needs an OMNIBOT_SESSION_TOKEN and an explicit --tab-id for the target tab.

How to click elements and fill forms with omnibot?▼

Start with semantic find commands such as find role button --name or find label --action fill, then fall back to snapshot -i refs like click @e4. Always verify the result with get, is, or wait after each action.

Does omnibot work with logged-in sessions and cookies?▼

Yes, omnibot drives the user's real visible browser, so existing login state, cookies, storage, and extensions are available. Headless mode does not inherit that login state, so use visible or background mode when authentication matters.

Why does my omnibot command fail with parser errors?▼

Parser failures usually come from remembered or outdated command syntax. Run omnibot --help and omnibot <command> --help to confirm current flags, especially for batch, network, cdp, record, trace, and skills commands.

When should I use execute-js instead of native omnibot commands?▼

Only as a last-resort fallback after semantic find, snapshot refs, selectors, DOM, and mouse tiers have failed with documented evidence. Native commands like get, read, click, fill, and scroll preserve event semantics and verification evidence that raw JavaScript bypasses.

Related Skills

Explore More Software Engineering Skills→
whiter001whiter001

automate-browser

Automate browser navigation, element interaction, and multi-tab workflows via v-browser CLI.

Community
Intermediate
kolodachkolodach

agent-browser

Automate browser interactions and extract page data via ref-based snapshots.

Community
Advanced
lucking7lucking7

browser-navigation

Automate browser interactions and extract data for testing and scraping.

Community
Intermediate
NagyViktNagyVikt
2

agent-browser

Automate browser interactions and capture screenshots for UI verification.

Community
Advanced