重要前提
安装AI Skills的关键前提是:必须科学上网,且开启TUN模式,这一点至关重要,直接决定安装能否顺利完成,在此郑重提醒三遍:科学上网,科学上网,科学上网。查看完整安装教程 →
BrightData by danielmiessler/personal_ai_infrastructure
npx skills add https://github.com/danielmiessler/personal_ai_infrastructure --skill BrightData执行前,请检查用户自定义配置位于: ~/.claude/skills/CORE/USER/SKILLCUSTOMIZATIONS/BrightData/
如果此目录存在,则加载并应用其中找到的任何 PREFERENCES.md、配置或资源。这些将覆盖默认行为。如果目录不存在,则使用技能默认设置。
执行工作流时,请同时执行以下两项操作:
发送语音通知:
curl -s -X POST http://localhost:8888/notify
-H "Content-Type: application/json"
-d '{"message": "Running the WORKFLOWNAME workflow from the BrightData skill"}'
> /dev/null 2>&1 &
输出文本通知:
Running the WorkflowName workflow from the BrightData skill...
完整文档: ~/.claude/skills/CORE/SkillNotifications.md
执行工作流时,直接输出此通知:
广告位招租
在这里展示您的产品或服务
触达数万 AI 开发者,精准高效
Running the **WorkflowName** workflow from the **Brightdata** skill...
关键:此工作流为 URL 内容检索实现了渐进式升级策略。
当用户请求抓取/获取 URL 内容时: 例如:"scrape this URL"、"fetch this page"、"get content from [URL]"、"pull content from this site"、"retrieve [URL]"、"can't access this site"、"this site is blocking me"、"use Bright Data to fetch" → 读取: ~/.claude/skills/brightdata/Workflows/Four-tier-scrape.md → 执行: 四级渐进式抓取工作流(WebFetch → Curl → 浏览器自动化 → Bright Data MCP)
渐进式升级策略:
主要特性:
four-tier-scrape.md - 使用四级回退策略进行完整的 URL 内容抓取
集成点:
mcp__Brightdata__scrape_as_markdown 用于高级抓取各级别的使用时机:
配置: 无需配置 - 所有工具在 Claude Code 中默认可用
示例 1:简单的公开网站
用户:"Scrape https://example.com"
技能响应:
示例 2:需要 JavaScript 的网站
用户:"Can't access this site https://dynamic-site.com"
技能响应:
示例 3:具有高级机器人检测的网站
用户:"Scrape https://protected-site.com"
技能响应:
示例 4:明确的 Bright Data 请求
用户:"Use Bright Data to fetch https://difficult-site.com"
技能响应:
相关文档:
~/.claude/skills/CORE/SkillSystem.md - 规范结构指南~/.claude/skills/CORE/CONSTITUTION.md - 整体 PAI 理念最后更新: 2025-11-23
每周安装次数
49
代码仓库
GitHub 星标数
9.7K
首次出现
Jan 24, 2026
安全审计
安装于
gemini-cli42
codex40
opencode39
github-copilot39
claude-code36
amp35
Before executing, check for user customizations at: ~/.claude/skills/CORE/USER/SKILLCUSTOMIZATIONS/BrightData/
If this directory exists, load and apply any PREFERENCES.md, configurations, or resources found there. These override default behavior. If the directory does not exist, proceed with skill defaults.
When executing a workflow, do BOTH:
Send voice notification :
curl -s -X POST http://localhost:8888/notify \
-H "Content-Type: application/json" \
-d '{"message": "Running the WORKFLOWNAME workflow from the BrightData skill"}' \
> /dev/null 2>&1 &
Output text notification :
Running the **WorkflowName** workflow from the **BrightData** skill...
Full documentation: ~/.claude/skills/CORE/SkillNotifications.md
When executing a workflow, output this notification directly:
Running the **WorkflowName** workflow from the **Brightdata** skill...
CRITICAL: This workflow implements progressive escalation for URL content retrieval.
When user requests scraping/fetching URL content: Examples: "scrape this URL", "fetch this page", "get content from [URL]", "pull content from this site", "retrieve [URL]", "can't access this site", "this site is blocking me", "use Bright Data to fetch" → READ: ~/.claude/skills/brightdata/Workflows/Four-tier-scrape.md → EXECUTE: Four-tier progressive scraping workflow (WebFetch → Curl → Browser Automation → Bright Data MCP)
Progressive Escalation Strategy:
Key Features:
four-tier-scrape.md - Complete URL content scraping with four-tier fallback strategy
Integration Points:
mcp__Brightdata__scrape_as_markdown for advanced scrapingWhen Each Tier Is Used:
Configuration: No configuration required - all tools are available by default in Claude Code
Example 1: Simple Public Website
User: "Scrape https://example.com"
Skill Response:
Example 2: Site with JavaScript Requirements
User: "Can't access this site https://dynamic-site.com"
Skill Response:
Example 3: Site with Advanced Bot Detection
User: "Scrape https://protected-site.com"
Skill Response:
Example 4: Explicit Bright Data Request
User: "Use Bright Data to fetch https://difficult-site.com"
Skill Response:
Related Documentation:
~/.claude/skills/CORE/SkillSystem.md - Canonical structure guide~/.claude/skills/CORE/CONSTITUTION.md - Overall PAI philosophyLast Updated: 2025-11-23
Weekly Installs
49
Repository
GitHub Stars
9.7K
First Seen
Jan 24, 2026
Security Audits
Gen Agent Trust HubFailSocketPassSnykWarn
Installed on
gemini-cli42
codex40
opencode39
github-copilot39
claude-code36
amp35
GitHub Actions 官方文档查询助手 - 精准解答 CI/CD 工作流问题
49,000 周安装