Browser

ruvnet/ruflo/v3/@claude-flow/cli/.claude/skills/browser

作者 ruvnet5f709e36799274be6bb68fff69090bdbc8621220无许可证74K 个星标收录于 2026年10月9日更新于 2026年10月9日仓库今天更新

Web browser automation with AI-optimized snapshots for claude-flow agents

AI 生成的概览

通过 agent-browser 命令行工具自动化网页浏览器,使用面向 AI 优化的可访问性快照和元素引用。

功能
该技能记录了一套驱动网页浏览器的命令行工作流:打开网址、获取可访问性树快照,并通过点击、填写表单、输入、滚动和等待与页面交互。内容涵盖导航、快照、交互、信息获取、等待、会话以及元素引用、CSS 选择器和语义定位器等选择器策略。它还说明了与 Claude Flow 的集成方式,包括 MCP 工具、记忆存储和钩子。
适用场景
当智能体需要浏览网站、填写并提交表单、提取页面数据或截取屏幕截图时使用。它也适用于需要隔离会话或共享认证状态的并行浏览任务。
运行要求
需要安装并可用 agent-browser 命令行工具,并具备访问目标站点的网络连接。可选的 Claude Flow 集成使用 npx @claude-flow/cli 进行记忆存储和钩子调用。该技能不附带脚本,仅为说明文档。

Browser Automation Skill

Web browser automation using agent-browser with AI-optimized snapshots. Reduces context by 93% using element refs (@e1, @e2) instead of full DOM.

Core Workflow

bash
# 1. Navigate to pageagent-browser open <url>
# 2. Get accessibility tree with element refsagent-browser snapshot -i    # -i = interactive elements only
# 3. Interact using refs from snapshotagent-browser click @e2agent-browser fill @e3 "text"
# 4. Re-snapshot after page changesagent-browser snapshot -i

Quick Reference

Navigation

CommandDescription
open <url>Navigate to URL
backGo back
forwardGo forward
reloadReload page
closeClose browser

Snapshots (AI-Optimized)

CommandDescription
snapshotFull accessibility tree
snapshot -iInteractive elements only (buttons, links, inputs)
snapshot -cCompact (remove empty elements)
snapshot -d 3Limit depth to 3 levels
screenshot [path]Capture screenshot (base64 if no path)

Interaction

CommandDescription
click <sel>Click element
fill <sel> <text>Clear and fill input
type <sel> <text>Type with key events
press <key>Press key (Enter, Tab, etc.)
hover <sel>Hover element
select <sel> <val>Select dropdown option
check/uncheck <sel>Toggle checkbox
scroll <dir> [px]Scroll page

Get Info

CommandDescription
get text <sel>Get text content
get html <sel>Get innerHTML
get value <sel>Get input value
get attr <sel> <attr>Get attribute
get titleGet page title
get urlGet current URL

Wait

CommandDescription
wait <selector>Wait for element
wait <ms>Wait milliseconds
wait --text "text"Wait for text
wait --url "pattern"Wait for URL
wait --load networkidleWait for load state

Sessions

CommandDescription
--session <name>Use isolated session
session listList active sessions

Selectors

Element Refs (Recommended)

bash
# Get refs from snapshotagent-browser snapshot -i# Output: button "Submit" [ref=e2]
# Use ref to interactagent-browser click @e2

CSS Selectors

bash
agent-browser click "#submit"agent-browser fill ".email-input" "[email protected]"

Semantic Locators

bash
agent-browser find role button click --name "Submit"agent-browser find label "Email" fill "[email protected]"agent-browser find testid "login-btn" click

Examples

Login Flow

bash
agent-browser open https://example.com/loginagent-browser snapshot -iagent-browser fill @e2 "[email protected]"agent-browser fill @e3 "password123"agent-browser click @e4agent-browser wait --url "**/dashboard"

Form Submission

bash
agent-browser open https://example.com/contactagent-browser snapshot -iagent-browser fill @e1 "John Doe"agent-browser fill @e2 "[email protected]"agent-browser fill @e3 "Hello, this is my message"agent-browser click @e4agent-browser wait --text "Thank you"

Data Extraction

bash
agent-browser open https://example.com/productsagent-browser snapshot -i# Iterate through product refsagent-browser get text @e1  # Product nameagent-browser get text @e2  # Priceagent-browser get attr @e3 href  # Link

Multi-Session (Swarm)

bash
# Session 1: Navigatoragent-browser --session nav open https://example.comagent-browser --session nav state save auth.json
# Session 2: Scraper (uses same auth)agent-browser --session scrape state load auth.jsonagent-browser --session scrape open https://example.com/dataagent-browser --session scrape snapshot -i

Integration with Claude Flow

MCP Tools

All browser operations are available as MCP tools with browser/ prefix:

  • browser/open
  • browser/snapshot
  • browser/click
  • browser/fill
  • browser/screenshot
  • etc.

Memory Integration

bash
# Store successful patternsnpx @claude-flow/cli memory store --namespace browser-patterns --key "login-flow" --value "snapshot->fill->click->wait"
# Retrieve before similar tasknpx @claude-flow/cli memory search --query "login automation"

Hooks

bash
# Pre-browse hook (get context)npx @claude-flow/cli hooks pre-edit --file "browser-task.ts"
# Post-browse hook (record success)npx @claude-flow/cli hooks post-task --task-id "browse-1" --success true

Tips

  1. Always use snapshots - They're optimized for AI with refs
  2. Prefer -i flag - Gets only interactive elements, smaller output
  3. Use refs, not selectors - More reliable, deterministic
  4. Re-snapshot after navigation - Page state changes
  5. Use sessions for parallel work - Each session is isolated

来源与署名

来源:ruvnet/ruflo位于v3/@claude-flow/cli/.claude/skills/browser提交5f709e3

许可证: 无许可证

内容归原作者所有。SourceWeft 从公开仓库中收录这些内容。

举报或申请下架