openclaw-skills/agent-browser/SKILL.md
Use when an agent needs real browser automation for semantic element targeting, form interaction, screenshots, recordings, script execution, session reuse, or end-to-end web workflow verification.
npx skillsauth add seaworld008/commonly-used-high-value-skills agent-browserInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
当任务需要 直接操作网页(而不是只读 API 数据)时使用本技能。Agent Browser 基于标准的 CDP (Chrome DevTools Protocol) 协议,赋予 Agent 像人类一样浏览网页、填写表单、点击按钮以及处理复杂前端交互(如 React/Vue 渲染页面、弹窗、验证码预览)的能力。
npx clawhub@latest install agent-browser
retain 机制),实现跨会话的免登录访问。Success Criteria。screenshot,既是留存证据,也方便 Agent 自我纠错。console 读取页面报错、性能指标(LCP, FID)并进行诊断。snapshot 获取辅助功能树(Accessibility Tree)。ax*(可访问性引用)和 n*(DOM 引用)理解页面层级。highlight 技能验证定位是否准确。scroll 滚动到底部,确保懒加载(Lazy Loading)的内容完整呈现。act 接口,执行 click, type, hover, drag 等动作。clearFirst: true 确保输入框内容干净。slowly: true 以模仿人类真实点击频率,降低被反爬引擎识别的风险。wait 工具,直到特定 URL 或文字出现,确保操作已生效。retain: true 标记重要的页签(Tab)。retain,等用户扫码完成后 Agent 再接管。errors 列表,识别 JS 执行崩溃。requests 网络请求,拦截并分析特定的 API 数据包(甚至包括 XHR/Fetch 的 Response)。### 任务目标
[目标描述:从 X 网站抓取前 10 个产品的价格和评论]
### 操作步骤 (Action Sequence)
1. **Navigate**: 访问 `https://example.com/products`。
2. **Wait**: 等待包含 `.product-list` 的元素加载完成。
3. **Scroll**: 滚动至页面底部触发分页。
4. **Snapshot**: 获取所有 `n` 级引用的文本内容。
5. **Console**: 注入脚本 `Array.from(document.querySelectorAll('.price')).map(e => e.innerText)`。
6. **Screenshot**: 截取整个列表区域存入 `output/products.png`。
### 预期结果
- 汇总 JSON 数据。
- 截图证据。
// 示例:打开 GitHub 并保留会话,不被自动销毁
const tab = await mcp_call({
name: 'browser_open',
arguments: {
targetUrl: "https://github.com",
profile: "chrome"
}
});
await mcp_call({
name: 'browser_retain',
arguments: {
targetId: tab.targetId,
retain: true
}
});
MEMORY.md 中的“基准截图”进行对比,识别 UI 错位或样式丢失。question 技能将截图发给用户)。timeoutMs。retain,否则在任务结束后及时 close 无效页签。incognito 模式。browser console 错误日志记录下来,它是调试的第一线索。development
Enumerating failure modes via pre-mortem analysis. Systematically identifies failure scenarios for plans, designs, and features, scoring them with RPN/AP. Does not write code.
testing
Orchestrating specialist AI agent teams as a meta-coordinator. Decomposes requests into minimum viable chains, spawns each as an independent session in AUTORUN modes, and drives to final output. Use when a task spans multiple specialist domains, requires parallel agent execution, or needs hub-and-spoke routing across the skill ecosystem.
development
Converting document formats (Markdown/Word/Excel/PDF/HTML). Converts specs from Scribe and reports from Harvest into distributable formats; generates reusable conversion scripts. Use when converting documents, building accessibility-compliant PDFs, or creating Pandoc/LibreOffice pipelines.
testing
Curating cross-agent knowledge and guarding institutional memory. Extracts patterns from agent journals into METAPATTERNS.md, detects knowledge decay, propagates best practices, prevents organizational forgetting. Use when consolidating cross-agent insights, curating memory, or auditing knowledge decay.