link-checker
Use when auditing webpages or URL lists for broken links, redirects, crawl health, HTTP status failures, timeouts, or link-report generation.
适合你，如果你需要定期检查网站链接的健康状况
npx oh-my-skill add qianleigood/crawclaw/link-checkercurl -fsSL https://oh-my-skill.com/install.sh | bash -s -- qianleigood/crawclaw/link-checkernpx oh-my-skill verify qianleigood/crawclaw/link-checker怎么用
商店整理自技能原文 · 版本 f27e175 · 表述以原文为准装好后，Claude 会变成链接检查器。你给一个网址，它就爬取页面检查所有超链接是否有错（如 404、重定向、超时），并给出详细报告。
当你提供网址并指定检查方式（如检查单页、找断链或递归爬取）时触发。
技能原文 SKILL.md
link-checker
Broken link finder that crawls web pages and checks all hyperlinks for HTTP errors (404 Not Found), redirects (301/302), timeouts, and connection failures. Generates detailed reports with status codes, response times, and link locations. Supports recursive crawling with depth control, domain filtering, and multiple output formats. Uses Python3 urllib — no external dependencies. Essential for website maintenance, SEO hygiene, and quality assurance.
Powered by BytesAgain | bytesagain.com | hello@bytesagain.com
Commands
| Command | Description | | ----------------- | --------------------------------- | | check <url> | Check all links on a single page | | crawl <url> | Recursively crawl and check links | | report <url> | Generate a full HTML report | | broken <url> | Show only broken links | | redirects <url> | Show only redirected links | | external <url> | Check only external links | | internal <url> | Check only internal links | | batch <file> | Check links on multiple pages | | summary <url> | Quick summary with counts |
Options
--depth <n>— Crawl depth for recursive mode (default: 1)--timeout <seconds>— Request timeout per link (default: 10)--format text|json|csv|html— Output format (default: text)--output <file>— Save report to file--concurrent <n>— Max concurrent checks (default: 5)--internal-only— Only check internal links--external-only— Only check external links--exclude <pattern>— Exclude URLs matching pattern--verbose— Show all links including OK ones--quiet— Only show broken links--user-agent <string>— Custom User-Agent header
Examples
# Check all links on a page bash scripts/main.sh check https://example.com # Find only broken links bash scripts/main.sh broken https://example.com --format json # Recursive crawl with depth limit bash scripts/main.sh crawl https://example.com --depth 2 --internal-only # Generate HTML report bash scripts/main.sh report https://example.com --output link-report.html # Check external links only with custom timeout bash scripts/main.sh external https://example.com --timeout 15 # Quick summary bash scripts/main.sh summary https://example.com
Status Categories
| Category | Status Codes | Description | | ----------- | ------------------ | ------------------------------ | | ✅ OK | 200, 204 | Link is working | | 🔄 Redirect | 301, 302, 307, 308 | Link redirects | | ❌ Broken | 404, 410 | Link is dead | | ⚠️ Error | 403, 500, 502, 503 | Server error | | ⏱️ Timeout | — | Connection timed out | | 🚫 Failed | — | Connection refused/DNS failure |
子代理执行策略
- 默认执行模式：主会话编排，子代理只负责读、查、预检、草拟、汇总等前后处理；真正的关键执行动作仍留在主会话。
- 子代理可做的事：读取现状、跑预检、整理参数、生成任务摘要、批量查询、拉取结果、汇总分析、给出推荐动作。
- 必须留在主会话的事：外部发送、远端写入、权限/配置修改、登录授权、绑定当前聊天、创建定时或持续运行任务、真实计费提交。
- 若技能自身已有后台机制（如 dispatcher / watcher / queue / monitor），子代理只做前后处理，不替代原生后台执行链路。
- 回报节点：默认保留“计划确认 / 预检结果 / 关键动作已执行 / 已完成 / 已失败”这类高价值节点。