A high-performance skill that leverages Cloudflare's protocol to fetch web content in clean Markdown format with significant token savings.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install cf-markdown-agents
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install cf-markdown-agents using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Cloudflare Markdown for Agents (cf-markdown-agents)?
This skill integrates Cloudflare's Markdown for Agents protocol into your AI workflow, enabling the automatic conversion of HTML pages into structured Markdown directly at the edge. By utilizing Openclaw Skills, developers can significantly reduce context window pressure when processing web data, as the protocol provides an average of 80% token reduction compared to raw HTML. This makes it an essential tool for building cost-effective and performant AI agents that need to consume web content reliably without the overhead of complex scraping logic.
The feature also handles metadata headers that are critical for modern AI operations, including token count estimation and usage permissions. It allows your agents to interact with web content in a format that is natively understood by LLMs, ensuring higher accuracy and lower latency during content synthesis or retrieval-augmented generation (RAG) tasks.
Cloudflare Markdown for Agents (cf-markdown-agents) Use Cases
- Reducing LLM token consumption by fetching web pages as optimized Markdown instead of raw HTML.
- Scouring developer documentation and blogs to build accurate AI training or RAG pipelines.
- Extracting clean, structured content from Cloudflare-powered websites without maintaining custom parsing scripts.
- Automated monitoring of web content changes while respecting AI-specific usage permissions via Content-Signal headers.
How Cloudflare Markdown for Agents (cf-markdown-agents) Works
- The agent or script initiates a request to a target URL using specific content negotiation headers (Accept: text/markdown).
- Cloudflare's edge network intercepts the request and identifies the agent's preference for Markdown.
- The HTML content of the target page is transformed into structured, clean Markdown before it ever reaches your agent.
- The skill captures the response, extracting the x-markdown-tokens header to provide an immediate estimate of the token cost.
- The finalized Markdown content is delivered to the AI model or database, ready for immediate processing with minimal noise.
Cloudflare Markdown for Agents (cf-markdown-agents) Setup
To start using this skill within your Openclaw Skills environment, you can use the provided fetch script or standard HTTP tools:
# Fetch a URL using the optimized script
scripts/fetch-markdown.sh "https://developers.cloudflare.com/agents/"
# Or use cURL directly
curl https://example.com/page -H "Accept: text/markdown"
Ensure your agent has the necessary network permissions to reach external Cloudflare-enabled domains.
Cloudflare Markdown for Agents (cf-markdown-agents) Data Schema & Taxonomy
The skill returns a structured response that prioritizes efficiency and metadata transparency. The following table describes the primary data points returned:
| Attribute | Description |
|---|---|
| content-type | The MIME type of the response, standardized as text/markdown; charset=utf-8 |
| x-markdown-tokens | An integer representing the estimated token count of the Markdown payload |
| content-signal | Metadata defining AI usage permissions (e.g., ai-train=yes, search=yes) |
| markdown_body | The processed, structured Markdown string of the target webpage |
Cloudflare Markdown for Agents (cf-markdown-agents) Advanced Features
- Automated token count tracking via x-markdown-tokens headers for precise budget and context window management.
- Built-in compatibility with Content-Signal headers to ensure your agent follows ethical data scraping and AI-usage guidelines.
- Zero-latency transformation at the edge, eliminating the need for heavy local parsing libraries like BeautifulSoup or Cheerio.
- Native support for Cloudflare's documentation ecosystem, providing the most reliable path for fetching technical reference data.
SKILL.md
Loading
Related Openclaw Skills
Coupon Finder
An intelligent discovery tool that matches user needs with optimal discount links across various lifestyle services.
LMP Label Generator
An intelligent design assistant that converts natural language into professional LMP label files for printing and digital preview.
Create Edu Video
An automated end-to-end educational video production skill using Remotion, TTS, and AI-driven scriptwriting.
Research Authorship & CRediT Generator
An AI-powered tool for generating fair authorship orders and CRediT contributor statements for research manuscripts.
Tencent Ecosystem Navigation
A specialized skill for navigating the Tencent ecosystem, including Tencent Cloud and WeChat, with region-aware guidance and official documentation verification.
Protocol Deviation Classifier
An automated tool for clinical trials that classifies protocol deviations based on GCP and ICH E6 guidelines to ensure data integrity and subject safety.
Featured*
