skills.rest
EliasOulkadi
EliasOulkadiCOMMUNITY
Explore all skills by this author →
𝕏@EliasOulkadiGitHub

agent-browser

Automate browser navigation, form filling, clicks, and evidence collection via accessibility-tree references.

110|14|Updated May 12, 2026
Search Tags:#screenshots#browser automation#session management#scraping#qa testing#accessibility tree#web form filling
One-click install
npx skills add https://github.com/EliasOulkadi/shokunin --skill agent-browser-eliasoulkadi
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/EliasOulkadi/shokunin/tree/main/.pack/skills/agent-browser
Command: npx skills add https://github.com/EliasOulkadi/shokunin --skill agent-browser-eliasoulkadi

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

agent-browser solves the problem of having to manually interact with complex web UIs (navigation, forms, clicks, and evidence collection) when you need accurate, repeatable browser automation for an AI agent.

Core Features & Use Cases

  • Accessibility-tree element targeting: Interact with pages using stable @eN references from accessibility snapshots to reduce brittle selectors.
  • Evidence and extraction tools: Capture screenshots, extract visible text/HTML, dump accessibility trees, view console and network logs, and generate PDFs.
  • Auth and session automation: Store encrypted credentials (vault) and persist browser state (sessions) to avoid repeated login flows.

Use it for QA and exploratory testing, form submission, login/OAuth flows, scraping data from a page, debugging UI issues, and automating Electron desktop apps like VS Code/Slack/Discord.

Quick Start

Tell your AI agent to navigate to your target site, wait for the needed element to appear using its accessibility reference, then click it and capture a screenshot as proof of the result.

Frequently Asked Questions about agent-browser

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

FAQPage Schema
How do I automate web form filling using accessibility tree references?▼

Browser automation uses accessibility-tree snapshots with element refs to interact with pages reliably. The agent identifies elements via `@eN` references from these snapshots, then executes clicks and form filling without relying on brittle CSS selectors.

Can I persist login sessions for browser automation to avoid repeated OAuth flows?▼

Yes, session automation persists browser state and stores encrypted credentials in a vault. This avoids repeated login or OAuth flows, maintaining authenticated states across multiple automation runs.

What is the best way to collect evidence during web QA testing?▼

The best way to collect QA evidence is capturing screenshots, extracting visible text or HTML, dumping accessibility trees, and inspecting console and network logs. This provides verifiable proof of the web page state during testing.

Does browser automation work with Electron desktop apps like VS Code or Slack?▼

Yes, browser automation supports Electron desktop app scenarios for VS Code, Slack, and Discord. It navigates the app UI, clicks controls, and extracts content using the same accessibility-tree snapshot mechanism as standard web pages.

Why use accessibility tree snapshots instead of CSS selectors for web scraping?▼

Accessibility tree snapshots provide stable `@eN` element references that reduce brittleness compared to CSS selectors. When scraping data or clicking controls, these stable references ensure interactions remain reliable even if the underlying HTML structure changes.

Related Skills

Explore More Software Engineering Skills→
adryanmoldokkr32-pixeladryanmoldokkr32-pixel

browse

Drive a Playwright-based headless browser to navigate, interact, and verify page states.

Community
Intermediate
carterwayneskhizeinecarterwayneskhizeine
19

ocr-and-documents

Extract text and structured content from PDFs and scanned documents.

Community
Intermediate
Spectaculous-CodeSpectaculous-Code

ci-doctor

Diagnose and resolve CI/CD pipeline and GitHub Actions failures.

Official
Advanced
SpaceZephyrSpaceZephyr
333

web-scraper

Fetch URLs and convert HTML page content into readable markdown.

Community
Intermediate