Token导航 LogoToken导航

Agent Skill · 研究检索

slbrow-skill(粗眉技巧)

slbrow-skill 用于处理浏览器自动化、网页检查和页面信息提取,适合在 OpenClaw 中需要让 Agent 打开页面、读取网页或验证前端流程时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

slbrow-skill 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为操作浏览器;许可证为 MIT-0,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 36,552。

总安装 1,905周安装 81GitHub Stars 公开资料未说明下载 667核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT-0
版本
1.0.0

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:slbrow-skill(粗眉技巧)
来源仓库:https://github.com/adeted/slbrow-skill
安装命令:
openclaw skills install slbrow-skill
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install slbrow-skill

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

slbrow-skill 用于浏览器自动化,支持页面导航、内容提取与选项卡管理。适合在 OpenClaw 中验证前端流程、抓取网页数据或模拟用户操作。通过 clawhub 安装,需连接 SLBrow MCP 或 HTTP API 服务。安装前应确认服务可用性、端口开放与认证方式。

涉及敏感页面时应遵守 robots.txt 与隐私政策限制。

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 OpenClaw;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 openclaw skills install slbrow-skill;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要操作浏览器,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

slbrow-skill 怎么安装?

在终端执行 openclaw skills install slbrow-skill。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

slbrow-skill 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

slbrow-skill 需要什么权限?

权限级别是操作浏览器。该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

slbrow-skill 是免费的吗?

来源标注的许可证是 MIT-0。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

slbrow-skill 用的人多吗?

公开数据是累计安装 1,905 次,近一周安装 81 次,下载 667 次。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

slbrow-skill 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

值得选的地方和要先想清楚的地方

值得选的地方

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。

先想清楚的地方

  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 OpenClaw,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
slbrow-skill

粗眉技巧 · 本页

1,905—操作浏览器MIT-0有提醒

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

name
slbrow-skills
description
Browser automation via SLBrow MCP and HTTP API. Use when the user needs to navigate pages, analyze content, manage tabs, search history, extract text, or control the browser. Supports both MCP tools and direct curl commands. Requires SLBrow server running and browser extension installed.

SLBrow Browser Automation (Skills for Browser)

Skills for browser-side AI assistant. The browser plugin (aiassist) connects to slbrow-mcp via MCP protocol; these skills guide when and how to use the tools. Can be used via MCP tools or direct HTTP API calls with curl.

Prerequisites

  1. SLBrow server running: npx slbrow or npm start in ai/slbrowmcp/slbrow-mcp/
  2. Browser extension installed and connected (Chrome/Firefox)
  3. MCP plugin enabled in aiassist: http://127.0.0.1:5556/mcp (视联浏览器MCP), http://127.0.0.1:5556/vaimcp (视联视频AI MCP)
  4. Default ports: HTTP 5556, WebSocket 5555

Usage Methods

Method 1: MCP Tools (Browser Plugin + MCP)

When the MCP plugin is enabled, the AI has access to browser tools. Use these skills to guide tool selection and workflow:

  • Navigation: Use page_navigate to open URLs
  • Analysis: Use page_analyze with intent_hint (article, form submit, post_create) before extraction
  • Tabs: Use tab_list first, then pass tab_id to target specific tabs

Method 2: HTTP API with curl (Direct Control)

Always available - Use curl commands to control browser directly without MCP plugin dependency:

Basic format (PowerShell-friendly — no temp JSON file):

curl.exe -X POST http://localhost:5556/api/execute -H 'Content-Type: application/json' -d '{\"tool\":\"TOOL_NAME\",\"args\":{\"PARAM\":\"VALUE\"}}'

Important notes for Windows PowerShell:

  • Use curl.exe instead of curl (PowerShell aliases curl to Invoke-WebRequest)
  • Put the JSON body in -d '...' with single quotes around the whole argument; backslash-escape every double quote in the JSON (\"). That way the shell passes valid JSON to curl without needing --data-binary @file

Common curl examples:

# Create new tab and navigate
curl.exe -X POST http://localhost:5556/api/execute -H 'Content-Type: application/json' -d '{\"tool\":\"tab_create\",\"args\":{\"url\":\"https://www.163.com\"}}'

# Navigate current tab
curl.exe -X POST http://localhost:5556/api/execute -H 'Content-Type: application/json' -d '{\"tool\":\"page_navigate\",\"args\":{\"url\":\"https://www.example.com\"}}'

# List all tabs
curl.exe -X POST http://localhost:5556/api/execute -H 'Content-Type: application/json' -d '{\"tool\":\"tab_list\",\"args\":{}}'

# Analyze page content
curl.exe -X POST http://localhost:5556/api/execute -H 'Content-Type: application/json' -d '{\"tool\":\"page_analyze\",\"args\":{\"intent_hint\":\"article\"}}'

# Extract article content
curl.exe -X POST http://localhost:5556/api/execute -H 'Content-Type: application/json' -d '{\"tool\":\"page_extract_content\",\"args\":{\"content_type\":\"article\"}}'

# Close specific tab
curl.exe -X POST http://localhost:5556/api/execute -H 'Content-Type: application/json' -d '{\"tool\":\"tab_close\",\"args\":{\"tab_id\":1234567890}}'

# Search browser history
curl.exe -X POST http://localhost:5556/api/execute -H 'Content-Type: application/json' -d '{\"tool\":\"get_history\",\"args\":{\"keywords\":\"search term\",\"max_results\":10}}'

Response format: API returns JSON with success, tool, result, and formatted fields for easy parsing.

Available Tools

ToolDescription
page_navigateNavigate current tab to URL
page_analyzeAnalyze page for elements (intent_hint: article, form submit, post_create, etc.)
page_extract_contentExtract content (content_type: article, search_results, posts)
page_wait_forWait for element or text (condition_type, selector, text)
tab_createCreate tab(s) (url, urls[], or count)
tab_closeClose tab(s) (tab_id, tab_ids[])
tab_listList open tabs with IDs
get_historySearch browser history (keywords, start_date, end_date, domains, max_results)
get_selected_textGet selected text (optional tab_id)
get_page_linksGet page links (link_type: all/internal/external, domains, max_results)

VAI Tools (Seelink)

ToolDescription
get_page_seelink_player_listGet all players on current page
use_seelink_players_aiApply AI function to players (ai_function_name: reduce_fog, face_mosaic, dark_reduce, human_outline, vechicle_outline, none)

Workflow Patterns

Multi-tab workflow:

  1. tab_list to get tab IDs
  2. Add tab_id to any tool to target a background tab

Content extraction:

  1. page_navigate to target URL
  2. page_analyze with intent_hint to find elements
  3. page_extract_content for article/search_results/posts

Form filling:

  1. page_analyze with intent_hint "form submit" to discover form elements
  2. Note the element IDs returned — use them with subsequent page_analyze or page_extract_content to confirm state

VAI (Seelink player AI):

  1. get_page_seelink_player_list to get player IDs and count
  2. Choose targets by player_position_list (zero-based index) or player_id_list (not both)
  3. use_seelink_players_ai with chosen targets and ai_function_name
  4. To apply different functions to different players, call use_seelink_players_ai separately for each
  5. To disable AI on a player, set ai_function_name to "none"

Fetching Skills in Browser

The browser plugin can fetch skills from the server to inject into AI context:

GET http://127.0.0.1:5556/api/skills

Returns the full SKILL.md content (text/markdown). Use when slbrow MCP plugins are enabled to guide tool usage.

API Reference

For full parameter schemas, see references/api-reference.md.

Additional resources

Error Handling

  • EXTENSION_DISCONNECTED: Ensure extension is installed and server is running. Check http://localhost:5556/health first.
  • Tool call timeout: Operation took >30s; retry with simpler args or break into smaller steps.
  • page_analyze returns 0 elements: Page may still be loading — wait a moment and retry, or page_navigate again.
  • tab_id invalid (tab was closed): Call tab_list to refresh available tab IDs before retrying.
  • Server restarted (session lost): Client will receive 404; re-initialize MCP connection (refresh the page).

Quick Reference for curl Commands

Tab Management:

  • Create tab: tab_create with url parameter
  • List tabs: tab_list with empty args
  • Close tab: tab_close with tab_id parameter
  • Navigate: page_navigate with url parameter

Content Analysis:

  • Analyze page: page_analyze with intent_hint (article, form_submit, post_create)
  • Extract content: page_extract_content with content_type (article, search_results, posts)
  • Get links: get_page_links with link_type (all, internal, external)

History & Text:

  • Search history: get_history with keywords, max_results
  • Get selected text: get_selected_text with optional tab_id

Seelink AI:

  • Get players: get_page_seelink_player_list
  • Apply AI: use_seelink_players_ai with player_position_list or player_id_list and ai_function_name

Troubleshooting

Server not responding:

  • Check if SLBrow server is running: curl.exe http://localhost:5556/health
  • Verify server is started with npx slbrow or npm start

Curl command fails:

  • Ensure using curl.exe in PowerShell
  • For -d, use single-quoted JSON with \" around each JSON string delimiter (see Method 2 examples above)
  • Verify server is accessible on port 5556

Tab operations not working:

  • Use tab_list to get current tab IDs first
  • Verify tab_id is valid and tab still exists
  • Check if browser extension is connected

平台分布

OpenClaw

88.05%
按下载量换算587

安全审计

VirusTotal

可疑

ClawScan

通过

Static analysis

可疑

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在提醒或未通过项,安装前请自行复核。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills