メニュー
CLI tool for AI agents to browse the web, interact with UI elements, and analyze frontend applications. Supports navigation, clicking, typing, and capturing snapshots for accessibility analysis.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | agent-browser |
| description | CLI tool for AI agents to browse the web, interact with UI elements, and analyze frontend applications. Supports navigation, clicking, typing, and capturing snapshots for accessibility analysis. |
agent-browser is a CLI tool designed for AI agents to interact with web pages. It allows you to navigate to URLs, inspect the accessibility tree (snapshot), interact with elements (click, type, fill), and extract information.
The typical workflow for analyzing and interacting with a page is:
agent-browser open <url>@12): agent-browser snapshot
--interactive or -i to see only interactive elements.agent-browser click @12agent-browser fill @15 "text"agent-browser get text @20agent-browser open <url> - Navigate to a URL.agent-browser back - Go back.agent-browser forward - Go forward.agent-browser reload - Reload the page.agent-browser snapshot - Returns the accessibility tree. This is the primary way to "see" the page structure and get element references.
-i (interactive only), -c (compact), -d <depth> (limit depth).agent-browser screenshot [path] - Take a screenshot (defaults to screenshot.png).agent-browser get <what> [selector] - Get details about an element.
what: text, html, value, attr <name>, title, url.agent-browser click <selector> - Click an element.agent-browser type <selector> <text> - Type text into an element.agent-browser fill <selector> <text> - Clear and fill an input.agent-browser press <key> - Press a key (e.g., Enter, Tab).agent-browser scroll <dir> [px] - Scroll up, down, left, right.agent-browser wait <selector|ms> - Wait for an element to appear or for a duration.agent-browser session list - List active sessions.--session <name> with commands to use a specific isolated session.1. Analyze a Landing Page
# Open the page
agent-browser open https://example.com
# Get the layout structure (interactive elements only for cleaner view)
agent-browser snapshot -i
# (Agent analyzes the output, e.g., sees a "Sign Up" button at ref @5)
2. Fill a Form
# Assuming we are on the form page and have identified elements via snapshot
agent-browser fill @10 "user@example.com"
agent-browser fill @12 "securepassword"
agent-browser click @15 # Submit button
3. Debugging/Verification
# Check if an element is visible
agent-browser is visible @error-message
# Take a screenshot to verify UI state
agent-browser screenshot error_state.png
Best practices for writing clean, maintainable, and robust tests with TUnit.
Use `helium-browser --dump-dom` to inspect the fully rendered DOM after JavaScript execution. For debugging layout/scrolling issues and scraping dynamic pages.
Senior UI/UX Engineer. Architect digital interfaces overriding default LLM biases. Enforces metric-based rules, strict component architecture, CSS hardware acceleration, and balanced design engineering.
Expert guide for creating, configuring, and managing Opencode Agents. Use this skill when the user wants to create a new agent or modify an existing one.
Expert guide and best practices for building production-ready web APIs with Axum 0.8+ in Rust. Use this skill when developing, refactoring, or structuring Axum applications.
Expert guide for testing Axum web applications. Use this skill when writing unit tests, integration tests, or setting up test infrastructure for Axum handlers, extractors, middleware, and full application testing. Covers tower::ServiceExt patterns, TestClient usage, mocking state/dependencies, and best practices.