Web Clipper to Notes

The web is where most people encounter new ideas, and it's also where most of those ideas die. The problem isn't capture — it's that raw web content doesn't fit a knowledge system. No frontmatter, no tags, no links to anything you already know. It's a foreign object in your vault. The skill makes web content native: structured note, source preserved cleanly, key ideas extracted as claims (not summaries), tags from your existing taxonomy, and suggested connections to notes you probably already have.

Effort gets matched to value via three tiers. Tier 1 (30 seconds) for reference content — recipes, documentation, how-tos. Clean markdown, frontmatter, source link, minimal extraction. Tier 2 (2 minutes) for articles and essays with ideas worth keeping — the default. Frontmatter plus key ideas as claims, two or three quotes only if they say something better than you could paraphrase, plus connection suggestions. Tier 3 (5 minutes) for foundational pieces — full literature note with idea cards, connection map, and integration suggestions. For batches of five or more URLs, the skill does a Tier-1 triage pass first, then asks which deserve deeper treatment. This prevents spending five minutes on something that turns out to be a duplicate of an article you clipped last month.

Frontmatter is opinionated and consistent. Title cleaned up. Author preserved (critical for credibility assessment later). Canonical URL with tracking parameters stripped — no utm_source, no fbclid, no Medium tracking garbage. Publish date if available, "unknown" if not (so you can find "articles I clipped without dates" later). Tags from your existing taxonomy, never invented on the fly — #psychology and #negotiation, not #cognitive-bias-anchoring-in-salary-negotiation. Granular tags help nobody; that's what wikilinks are for.

Two principles deliberately stay rigid. The My Reaction section is always present and always blank — it's the prompt that turns a clip into knowledge, and pre-filling it removes the only thing that forces you to actually think about what you saved. And the skill is honest about content quality: if an article is mostly fluff with one good paragraph, you get "one idea worth keeping from an otherwise thin article" — not manufactured value to justify the clip. A vault full of inflated clip notes is just a bookmarks folder with extra steps.