Catch broken Markdown links before they reach readers, using a lightweight CLI designed for reliable documentation checks in Openclaw Skills workflows.
安装与下载
1. ClawHub CLI
从源直接安装技能的最快方式。
npx clawhub@latest install md-link-audit
2. 手动安装
将技能文件夹复制到以下位置之一
~/.openclaw/skills/ <project>/skills/ 优先级：工作区 > 本地 > 内置
3. 提示词安装
将此提示词复制到 OpenClaw 即可自动安装。
请帮我使用 Clawhub 安装 md-link-audit。如果尚未安装 Clawhub，请先安装（npm i -g clawhub）。
更倾向于下载？
获取包含原生技能文件的ZIP压缩包。
什么是 Markdown Link Auditor？
Markdown Link Auditor is a zero-dependency Python CLI for checking links in README files, wikis, and technical manuals. Use it before publishing or after a project path refactor to identify dead links and references that may have drifted.
For CI/CD, it provides deterministic results, machine-readable --json output, and return codes that distinguish a clean result, a warning, and bad input. Its CI-friendly approach is intended to reduce misleading failures associated with external-link timeouts, making it a practical addition to Openclaw Skills for documentation maintenance.
Markdown Link Auditor 应用场景
- Check a README, wiki, or technical manual for broken links before publishing.
- Add a Markdown link audit to CI/CD and use return codes to control pass thresholds.
- Review documentation references after moving or renaming project paths.
- Pipe Markdown through standard input when the document is supplied by another command.
- Use JSON output when a script or CI job needs to consume audit results.
Markdown Link Auditor 工作原理
- Provide a Markdown file path or send Markdown content through standard input.
- Run the Python auditor, adding
--jsonwhen machine-readable output is needed. - Review the reported link status and any error details.
- Check the return code:
0indicates success,1indicates a warning, and2indicates bad input. - Fix or otherwise address reported dead paths, then rerun the audit to verify the result.
Markdown Link Auditor 配置指南
Requirements: Python 3.8 or newer. The skill declares no Python package dependencies. Run the script from the skill directory or adjust the path to match your checkout.
python3 scripts/md_link_audit.py README.md
To read Markdown from standard input:
cat README.md | python3 scripts/md_link_audit.py
For machine-readable output:
python3 scripts/md_link_audit.py README.md --json
Use the resulting exit code in CI to define how warnings affect your pipeline. Openclaw Skills users can invoke the CLI locally or from an existing build job; the provided skill documentation does not specify a separate package installation step.
Markdown Link Auditor 数据架构与分类体系
The auditor reports results to standard output and does not document generated report files. With --json, output is described as a structured JSON intermediate representation.
| Output element | Meaning |
|---|---|
result |
Audit result payload |
rc |
Result status code: 0 for success, 1 for warning, 2 for bad input |
error_code |
Error classification when applicable |
warnings |
Warning information when applicable |
The skill documentation does not define the exact types or nested fields for each JSON element. No separate metadata taxonomy or on-disk data format is specified.
Markdown Link Auditor 高级特性
- Machine-readable mode:
--jsonexposes structured results for scripts and CI integrations. - Pipeline input: Accept Markdown through standard input as well as a file argument.
- Deterministic CI status: Return codes
0,1, and2let teams distinguish success, warnings, and invalid input. - Lightweight runtime: Requires Python 3.8+ and declares no third-party Python packages.
- Cross-platform CLI: Metadata lists Windows, macOS, and Linux compatibility.
- External-timeout awareness: Described as CI-friendly for avoiding false positives caused by external-network timeouts; the skill metadata declares network access as disabled, so confirm the implementation's remote-link behavior for workflows that require live URL checks.
SKILL.md
加载中
Openclaw 相关技能
UDI Database Submission Compliance Checker
Determine which UDI database to submit to, when to submit, and which fields are required—with structured warnings when information is incomplete.
Clinical Evaluation Route Picker
Clinical Evaluation Route Picker helps teams assess clinical evaluation routes early and returns a deterministic recommendation with a machine-readable decision code.