Clip Resource

Bold terms in this file are defined in GLOSSARY.md.

For the case where the user has no Web Clipper, no Readwise, nothing installed: this skill fetches a URL and publishes a Resource file that looks like Web Clipper's own output. Each URL runs as independent recoverable Resource work under /using-slipbox; extraction and the template-resolved draft are checkpointed before the frozen target is published. It never touches .slipbox/candidates/ or takes part in the candidate pipeline; it reads template paths via .slipbox/bin/slipbox config get templates.<type>_path, filename/frontmatter conventions from the configured template and established vault rules, and transcript languages via .slipbox/bin/slipbox config get transcript_languages, not pipeline bookkeeping. make-literature-note (and other note-writing skills) read the Resource file later; this skill's job ends once it's published.

Prerequisite

• MUST: .slipbox/AGENTS.md exists — confirms setup-slipbox ran to completion.
• MUST: All dependencies (Defuddle, Firecrawl, youtube-transcript-api, TinyFish) are available — check at step-entry, not here.
• NEVER: proceed without either. Stop and tell the user to run setup-slipbox first.
• NEVER: improvise a naming/folder convention in place of running setup-slipbox.
• NEVER: install a missing dependency yourself — that's setup-slipbox's job. Point the user there instead.

Workflow

01 - Take the URL(s)

Ask for one or more URLs: articles, news stories, social/forum threads, or video links, any mix. There is no paste-the-text fallback. If the user hands you raw text instead of a link, tell them this skill only takes URLs and ask for one.