A comprehensive skill for using agent-browser, a CLI tool for browser automation designed for AI agents, developed by Vercel Labs. This skill covers installation, core commands, selectors (refs, CSS, XPath, semantic locators), agent mode, sessions, options, and best practices. Use this skill whenever the user needs to automate browser interactions via CLI commands, especially for AI agents that need to interact with web pages.
Use this skill whenever the user wants to:
This skill is organized to match the agent-browser official documentation structure (https://github.com/vercel-labs/agent-browser/blob/main/README.md). When working with agent-browser:
Install agent-browser:
examples/getting-started/installation.md for installation instructionsQuick Start:
examples/quick-start/quick-start.md for basic workflow examplesLearn core commands:
examples/commands/basic-commands.md for basic commands (open, click, fill, etc.)examples/commands/advanced-commands.md for advanced commands (snapshot, eval, etc.)examples/commands/get-info/ for information retrieval commandsexamples/commands/check-state/ for state checking commandsexamples/commands/find-elements/ for semantic locator commandsexamples/commands/wait/ for wait commandsexamples/commands/mouse-control/ for mouse control commandsexamples/commands/browser-settings/ for browser configurationexamples/commands/cookies-storage/ for cookies and storage managementexamples/commands/network/ for network interceptionexamples/commands/tabs-windows/ for tab and window managementexamples/commands/frames/ for iframe handlingexamples/commands/dialogs/ for dialog handlingexamples/commands/debug/ for debugging commandsexamples/commands/navigation/ for navigation commandsexamples/commands/setup/ for setup commandsUnderstand selectors:
examples/selectors/refs.md for refs-based selection (@e1, @e2, etc.)examples/selectors/traditional-selectors.md for CSS, XPath, and semantic locatorsUse agent mode:
examples/agent-mode/introduction.md for agent mode overviewexamples/agent-mode/optimal-workflow.md for optimal AI workflowexamples/agent-mode/integration.md for integrating with AI agentsAdvanced features:
examples/advanced/sessions.md for session managementexamples/advanced/headed-mode.md for debugging with visible browserexamples/advanced/authenticated-sessions.md for authentication via headersexamples/advanced/custom-executable.md for custom browser executableexamples/advanced/cdp-mode.md for Chrome DevTools Protocol integrationexamples/advanced/streaming.md for browser viewport streamingexamples/advanced/architecture.md for architecture overviewexamples/advanced/platforms.md for platform supportexamples/advanced/usage-with-agents.md for AI agent integration patternsConfigure options:
examples/options/global-options.md for global CLI optionsexamples/options/snapshot-options.md for snapshot-specific optionsexamples/options/session-options.md for session management optionsReference API documentation when needed:
api/commands.md - Complete command referenceapi/selectors.md - Selector referenceapi/options.md - Options referenceUse templates for quick start:
templates/basic-automation.md - Basic automation workflowtemplates/ai-agent-workflow.md - AI agent workflow templateThis skill includes detailed examples organized to match the official documentation structure. All examples are in the examples/ directory (see mapping above).
To use examples:
To use templates:
templates/ directory for common scaffoldingapi/commands.md - Complete command reference with syntax and examplesapi/selectors.md - Selector types and usage referenceapi/options.md - All options reference--json flag for machine-readable output in agent mode--session to maintain state across commands-i flag for interactive snapshot selection--headed for debugging, headless for production--cdp for Chrome DevTools Protocol integrationAGENT_BROWSER_STREAM_PORT for live browser preview--headers for authentication without login flows--executable-path for serverless deployments or custom browsers-i, -c, -d, -s options to optimize snapshot outputagent-browser, CLI browser automation, AI agents, browser automation CLI, refs, snapshot, agent mode, semantic locators, browser automation tool, command-line browser, AI agent browser, deterministic selectors, accessibility tree, browser commands, web automation CLI, sessions, headed mode, authenticated sessions, CDP mode, streaming, Chrome DevTools Protocol, Playwright, browser automation for AI
确保开发环境已安装必要的依赖和工具。
根据项目需求进行基础配置。
按照示例代码实现核心功能。
运行测试确保功能正常。
完成开发后进行部署和监控。