Harvest

Take one piece of content and ask a single question: is there anything in here worth pulling into LifeOS? Read it, hold every candidate idea against what the system already is and already has, and report what is worth adopting, where it maps, and how we would use it. System adoption is report-only — changing the system is a separate, approved step. Then, always as the final step, ingest the content into the Knowledge Archive via the _HARVEST pipeline so the source itself is preserved.

Workflow Routing

Quick Reference

• One input at a time. A URL, a YouTube link, an article, a raw paste, or a file path.
• Fetch first, then judge. YouTube → fabric -y <url> for the transcript. Article → WebFetch or Research. File → Read. Raw text → use as-is.
• Judge against the real system, not a guess of it. Map each candidate to a concrete surface: an Algorithm phase or gate, a hook, a specific skill, the memory system, Pulse, routing/EFFORT_MODEL, ISA, or a doctrine file. If you can't name the surface, it isn't a harvest hit yet.
• Prior Status is the honesty check. NEW / PARTIAL (we have something close) / DONE (already implemented) / REJECTED (considered and declined). Cross-check before calling anything NEW.
• Output is a ranked table plus an honest verdict. Rank by usefulness × novelty × effort. If nothing is worth adopting, say so plainly — a clean "nothing here" beats manufactured findings.

Relationship to neighbors