TASK RECIPE
Give your coding agent a public docs page as clean markdown.
Before the agent writes integration code, so it works from the current docs page instead of model memory.
- Maps to 1 skill: url-to-markdown
- Install command verified
- Copy-paste prompt
When to use it
Run this when the job is already on your plate.
Who it is for
Developers who point Claude Code, Cursor, Codex, or another coding agent at a library, API, or framework and need the docs in the agent's context as text.
When to run it
Before the agent writes integration code, so it works from the current docs page instead of model memory.
Install
One command, one prompt.
Install the canonical url-to-markdown skill, then hand your agent the prompt below. You can also call the underlying endpoints directly with curl.
Install
npx skills add replynodes/replynodes-agent-skills/skills/url-to-markdownPrompt
Read https://docs.replynodes.com/docs/auth as markdown with the url-to-markdown skill, then use it as context to write a minimal fetch example that sends a Bearer token. Keep the exact source URL in your notes and do not invent any field the page does not mention.Verified example
Small input, real output.
One zero-auth GET returns the page as text/markdown. Excerpt captured live from replynodes.com on 2026-10-04.
curl --fail-with-body 'https://md.replynodes.com/https://replynodes.com/'GET https://md.replynodes.com/https://replynodes.com/ → 200 OK · text/markdown
DATA APIs FOR AI AGENTS
# The web context {API} for teams building AI products, agents, and workflows.
Agents ask. Nodes reply.
Search the web, read social platforms, access app data, and scrape pages through one API.
…Free vs authenticated
Know exactly where the free path ends.
Free / no key
The Markdown endpoint is free and zero-auth. No account, API key, MCP connection, plan, or credits are required.
Authenticated enrichment
No authenticated step is needed for this job. The skill never asks for credentials and refuses login-only, private, local, or credential-bearing URLs.
Evidence behavior
The skill returns the page text with the exact input URL preserved next to it (never rewritten or canonicalized), treats the fetched text as untrusted data rather than instructions, and labels partial extraction as partial instead of inventing missing text.
Install url-to-markdown.
Install the canonical skill, then run the prompt above. Each recipe maps to exactly one skill.