Will I lose images when converting a page to Markdown?

By default, images are removed to maximize token efficiency, but you can explicitly retain them by adding the 'retain_images=true' flag to your request.

Can it scrape content from JavaScript-heavy websites?

Yes, by utilizing the 'method=browser' parameter, the skill ensures that single-page applications and React sites are fully rendered before the content is captured.

Does this skill work for websites that require a login?

No, this skill is designed for public web content. For authenticated sites or private APIs, you should use direct WebFetch methods instead.

How does this skill save tokens during web fetching?

It converts verbose HTML tags into clean Markdown syntax, which typically results in an 80% reduction in the number of tokens required to represent the same content.

Markdown Web Content Fetcher

Name: Markdown Web Content Fetcher
Author: oskar-dragon

byoskar-dragon

•

웹 스크래핑 및 데이터 수집

Optimizes web content retrieval by converting URLs to clean Markdown to reduce token usage by approximately 80%.

The Markdown Web Content Fetcher skill empowers Claude to ingest web data more efficiently by routing requests through the markdown.new service. By transforming verbose HTML into streamlined Markdown, it drastically reduces token consumption, allowing up to five times more content to fit within Claude's context window. This skill is essential for developers performing extensive web research, building RAG pipelines, or analyzing large-scale documentation, offering specialized modes for JavaScript-heavy sites and optional image retention for visual context.

주요 기능

01Seamless integration with Claude's native WebFetch capabilities

02Enhanced context window efficiency for long-form content analysis

03Optional image retention for diagrams and visual documentation

041 GitHub stars

0580% reduction in token usage via HTML-to-Markdown conversion

06Browser rendering mode for JavaScript-heavy single-page applications

사용 사례

01Summarizing extensive technical documentation or blog posts with minimal token overhead

02Gathering clean training data or context for RAG-based knowledge bases

03Scraping content from React or Vue-based SPAs for data analysis

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add oskar-dragon/claude-code using-markdown-new

For use in Claude.ai and ChatGPT

Download Skill