Daily Note to Knowledge

The daily note is where most knowledge systems begin and where most of them quietly stall. People capture because something felt important in the moment — a realisation in a meeting, a pattern they half-noticed, a line from a book — but capture without processing is just journalling. Zettelkasten works because it forces two transformations: translating an observation into a standalone idea, and connecting that idea to what you already know. Both transformations require thinking. The notes are the residue of that thinking, not a substitute for it.

The skill reads the whole day's note before extracting anything, because the day's context matters — what the user was doing, what was on their mind. Then it scans for the eight things that become permanent-note candidates: realisations, pattern observations, mental models, decision rationales, contradictions, investigable questions, connections, and operating-condition data ("I always feel drained after X"). Everything else stays in the daily log where it belongs. Logistics aren't insights, and pretending they are bloats the vault.

Three filters then run on every candidate. The 6-month test — will this be useful or interesting in six months, stripped of today's context? "Follow up with Sarah" fails; "design reviews work better when aesthetic and functional feedback are separated" passes. The atomic test — can this be expressed as a single, specific claim? If it needs three paragraphs and five caveats, it's either not ready or it's actually three notes. The novelty test — is this genuinely new, or is the user rediscovering something they already noted? Rediscoveries get flagged as additions to the existing note, not duplicates.

Each surviving idea becomes a permanent note written in the user's voice (not cleaned up into AI prose), with a one-sentence "why this matters", two-to-five wikilinks each carrying a specific rationale ("connects to X because the mechanism is the same"), and an Open Thread pointing forward. When a day is purely operational, the skill says so — "no permanent knowledge to extract, that's normal." Not every day produces lasting insight. A productive day and a knowledge-producing day are different things, and treating them as the same is what kills the practice.