mdnew
thendcn 제작
Web & Frontend Development요약
Fetch clean, agent-optimized Markdown from any URL using the markdown.new service.
설명
이 스킬이 하는 일
Fetches clean, token-efficient Markdown from any URL using the markdown.new three-tier conversion pipeline. Strips boilerplate, ads, and navigation menus, leaving only core content. Handles JS-heavy pages via Cloudflare Browser Rendering as a fallback.
Reduces page size by up to 80% compared to raw HTML, making it practical for agents with tight context windows that would otherwise struggle with full-page content.
사용 시점
- Extracting article text for LLM analysis
- Scraping docs pages when web_fetch returns noisy HTML
- Reading JS-rendered pages without a full browser setup
- Reducing token usage when ingesting long web pages
- Feeding clean web content into agent context windows
예시 워크플로
User asks: Summarize the content at https://docs.example.com/guide
- 1
Run python3 scripts/mdnew.py https://docs.example.com/guide
- 2
Pipeline attempts Header Negotiation first, falls back to Workers AI, then Cloudflare Browser Rendering
- 3
Receive clean Markdown with x-markdown-tokens count in the response
- 4
Pass Markdown to LLM for summarization
Clean Markdown summary of the documentation page, using a fraction of the tokens raw HTML would require
설치
clawhub install thendcn/mdnewCLI 명령을 실행하거나, 프롬프트를 Claude Code, Cursor, Codex 등 사용하는 코딩 에이전트에 붙여넣으세요.