菜单
Headless browser automation CLI for AI agents. Use for web scraping, testing, form filling, screenshots, and any browser interaction. Triggers: "browse", "open website", "click button", "fill form", "screenshot", "scrape", "test login flow", "automate browser", "web automation".
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | agent-browser |
| description | Headless browser automation CLI for AI agents. Use for web scraping, testing, form filling, screenshots, and any browser interaction. Triggers: "browse", "open website", "click button", "fill form", "screenshot", "scrape", "test login flow", "automate browser", "web automation". |
CLI browser automation optimized for AI. Use refs from snapshots for reliable element selection. 93% less context than Playwright MCP, 95% first-try success rate.
agent-browser open <url> # Navigate
agent-browser snapshot -i # Get interactive elements with refs
agent-browser click @e2 # Click by ref
agent-browser fill @e3 "text" # Fill by ref
agent-browser snapshot -i # Re-snapshot after changes
agent-browser close # Close browser
Get accessibility tree with element refs for deterministic selection:
agent-browser snapshot # Full tree
agent-browser snapshot -i # Interactive only (buttons, inputs, links)
agent-browser snapshot -c # Compact (no empty elements)
agent-browser snapshot -d 3 # Limit depth
agent-browser snapshot -i -c --json # Combine for AI parsing
Output example:
- heading "Example Domain" [ref=e1]
- button "Submit" [ref=e2]
- textbox "Email" [ref=e3]
Use refs with @ prefix: agent-browser click @e2
agent-browser open <url> # Go to URL
agent-browser back # Go back
agent-browser forward # Go forward
agent-browser reload # Reload page
agent-browser click @ref # Click element
agent-browser fill @ref "text" # Clear and fill input
agent-browser type @ref "text" # Type into element
agent-browser press Enter # Press key
agent-browser select @ref "value" # Select dropdown option
agent-browser check @ref # Check checkbox
agent-browser hover @ref # Hover element
agent-browser scroll down 500 # Scroll (up/down/left/right)
agent-browser get text @ref # Get text content
agent-browser get value @ref # Get input value
agent-browser get title # Page title
agent-browser get url # Current URL
agent-browser is visible @ref # Check visibility
agent-browser is enabled @ref # Check if enabled
agent-browser is checked @ref # Check if checked
agent-browser wait @ref # Wait for element visible
agent-browser wait 2000 # Wait milliseconds
agent-browser wait --text "Welcome" # Wait for text
agent-browser wait --load networkidle # Wait for network idle
agent-browser screenshot # Base64 to stdout
agent-browser screenshot page.png # Save to file
agent-browser screenshot --full # Full page
agent-browser pdf output.pdf # Save as PDF
agent-browser open https://app.example.com/login
agent-browser snapshot -i
# Identify: textbox "Email" [ref=e1], textbox "Password" [ref=e2], button "Sign In" [ref=e3]
agent-browser fill @e1 "test@example.com"
agent-browser fill @e2 "password123"
agent-browser click @e3
agent-browser wait --url "**/dashboard"
agent-browser snapshot -i # Verify dashboard loaded
agent-browser get text @e1 # Verify welcome message
agent-browser open https://app.example.com/contact
agent-browser snapshot -i
agent-browser fill @e1 "John Doe" # Name field
agent-browser fill @e2 "john@example.com" # Email field
agent-browser fill @e3 "Hello world" # Message field
agent-browser click @e4 # Submit button
agent-browser wait --text "Thank you" # Verify success message
agent-browser snapshot -i # Confirm state
# Save auth after login
agent-browser open https://app.example.com/login
agent-browser snapshot -i && agent-browser fill @e1 "user@test.com"
agent-browser fill @e2 "password" && agent-browser click @e3
agent-browser wait --load networkidle
agent-browser state save auth.json # Save cookies/storage
# Reuse in later tests
agent-browser state load auth.json # Load saved state
agent-browser open https://app.example.com/dashboard # Already logged in
# Test complete user journey
agent-browser open https://shop.example.com
agent-browser snapshot -i
# Search for product
agent-browser fill @e1 "laptop" # Search box
agent-browser click @e2 # Search button
agent-browser wait --load networkidle
agent-browser snapshot -i
# Add to cart
agent-browser click @e5 # First product
agent-browser wait --load networkidle
agent-browser snapshot -i
agent-browser click @e3 # Add to cart button
agent-browser wait --text "Added to cart"
# Checkout
agent-browser click @e4 # Cart icon
agent-browser snapshot -i
agent-browser click @e2 # Checkout button
agent-browser wait --url "**/checkout"
agent-browser open https://app.example.com/component
agent-browser wait --load networkidle
agent-browser screenshot baseline.png --full
# Make changes, then compare
agent-browser screenshot current.png --full
agent-browser open https://app.example.com/form
agent-browser snapshot -i
agent-browser click @e5 # Submit without filling required fields
agent-browser wait 500
agent-browser snapshot -i # Capture error states
agent-browser get text @e1 # Verify error message text
# Run tests in isolated sessions
agent-browser --session test1 open https://app.example.com/feature-a
agent-browser --session test2 open https://app.example.com/feature-b
# Each session has separate cookies/storage
agent-browser --session test1 snapshot -i
agent-browser --session test2 snapshot -i
# List active sessions
agent-browser session list
agent-browser open url --headed # See browser window
agent-browser console # View console messages
agent-browser errors # View page errors
agent-browser highlight @e3 # Highlight element visually
agent-browser trace start # Record trace
# ... run test steps ...
agent-browser trace stop trace.zip # Save for analysis
When refs aren't available:
agent-browser click "#submit"
agent-browser click ".btn-primary"
agent-browser click "text=Sign In"
agent-browser find role button click --name "Submit"
| Flag | Description |
|---|---|
--session <name> | Use isolated session |
--json | JSON output (for parsing) |
--headed | Show browser window |
--full | Full page screenshot |
wait after actions - Ensure page state settles--json for assertions - Parse structured output--headed for debugging - See what's happeningProfile a repository, inspect existing Litho/deepwiki-rs cache state, generate a repo-local Litho config that preserves the user's current global model/provider, and run deepwiki-rs within a target time budget. Use when Claude needs to analyze a repo quickly with Litho, auto-tune exclusions and boundary settings by repo size, reuse cache intelligently, or create a filtered local copy for very large repos.
Create OmoiOS Improvement Proposals (OIPs) through an interactive workflow. Gathers context from the user via clarifying questions, researches relevant codebase files, then writes a complete proposal following the OIP template. Use when requesting "create a proposal", "write an OIP", "new proposal", "propose a feature", "OIP", "improvement proposal", or when the user wants to formally propose a product, architecture, or process change for OmoiOS.
Spec-driven development workflow for turning feature ideas into structured PRDs, requirements, designs, tickets, and tasks. Uses a state machine approach with EXPLORE → REQUIREMENTS → DESIGN → TASKS → SYNC phases. Each phase has validation gates, checkpointing, and session transcript support for cross-sandbox resumption.
Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, or extract information from web pages.
Spec-driven development workflow for turning feature ideas into structured PRDs, requirements, designs, tickets, and tasks. Uses a state machine approach with EXPLORE → REQUIREMENTS → DESIGN → TASKS → SYNC phases. Each phase has validation gates, checkpointing, and session transcript support for cross-sandbox resumption.
Build automated billing systems for recurring payments, invoicing, subscription lifecycle, and dunning management. Use when implementing subscription billing, automating invoicing, or managing recurring payment systems.