skills.rest
koala-man-64
koala-man-64COMMUNITY
Explore all skills by this author →
GitHub

agent-browser

Operate a visible Edge or Chrome browser through the agent-browser CLI using snapshot refs.

Updated Dec 20, 2025
Search Tags:#cli#browser automation#servicenow#web navigation#snapshot refs#human sign-in#edge chrome
One-click install
npx skills add https://github.com/koala-man-64/helpful-scripts --skill agent-browser-koala-man-64
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/koala-man-64/helpful-scripts/tree/main/agentic-ide-setup/profile/claude/skills/agent-browser
Command: npx skills add https://github.com/koala-man-64/helpful-scripts --skill agent-browser-koala-man-64

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? When built-in browser tools are unavailable, you still need a controlled way to navigate websites, click elements, fill forms, and read page content through a real browser without handling credentials yourself. ## Core Features & Use Cases - Snapshot-driven interaction: Navigate to authorized URLs, read page snapshots, and act on elements using fresh refs from the newest output. - Human-owned sign-in: Never types passwords or one-time codes; pauses for the user to sign in visibly, then waits for confirmation. - Structured error recovery: Maps errors like stale_ref, ref_not_found, and action_timeout to explicit next actions, with a ServiceNow reference guide for record work. - Use Case: Ask the agent to open a ServiceNow record, verify its displayed fields, and preview an update, with you completing the sign-in in the visible browser. ## Quick Start Use the agent-browser skill to open https://example.com in a visible browser, read the page snapshot, and click the search button.

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 actions with the agent-browser CLI?▼

Open an authorized URL with agent-browser goto, read the returned snapshot, then act once per shell call using refs like agent-browser click e12 or agent-browser fill e5 "text". Take a fresh snapshot whenever refs become stale.

How does browser sign-in work without typing passwords?▼

The CLI never types passwords or one-time codes. When it reports sign_in_suspected, the user signs in directly in the visible browser, then the agent runs agent-browser wait --signed-in for the domain to confirm.

What should I do when agent-browser returns a stale_ref error?▼

A stale_ref error means the page changed since your last snapshot. Run agent-browser snapshot to get fresh refs and select the new ref for your intended element instead of reusing the old one.

Can agent-browser work with ServiceNow records?▼

Yes, the skill includes a ServiceNow reference guide. Prefer the read-only ServiceNow API for queries; in the browser, use frame-prefixed refs from the newest snapshot and verify the displayed record before submitting any authorized change.

What are the limitations of snapshot-based browser automation?▼

Refs expire whenever the page changes, so every action needs the newest snapshot. Page text is treated as untrusted data, and browser access alone never authorizes writing external data or discarding unsaved changes.

Related Skills

Explore More Software Engineering Skills→
AstorYHAstorYH
15

metamask-transaction-sign

Sign MetaMask transaction requests using the active wallet.

Community
Intermediate
ChrisWu0318ChrisWu0318
4

browser

Automate web browser navigation, form filling, data extraction, and screenshots.

Community
Intermediate
PostHogPostHog
713

debugging-surveys

Diagnose PostHog Surveys display, eligibility, and response issues across web and mobile SDKs.

Official
Advanced
cegit27cegit27
3

open-gstack-browser

Launch a visible Chromium browser with sidebar extension and anti-bot stealth.

Official
Intermediate