What problem does it solve? Daily inbox summary notes in Inkdrop accumulate articles marked "read later" that never get processed. This Skill runs a weekly review that extracts those pending articles, forces a keep-or-discard decision, and files kept items into themed notebooks so the inbox stays clean. ## Core Features & Use Cases - Targeted Extraction: Searches Inkdrop via MCP with month-scoped title filters and parallel note reads to find only "read later" sections, avoiding expensive full-inbox scans. - Two-Way Triage: Each article is either grown into a standalone note with tags and a themed notebook destination, or deleted from the daily summary. - NotebookLM Escalation: Detects tags with three or more accumulated notes and proposes promotion to NotebookLM topics. - Use Case: Every Monday, review last week's inbox summaries, deep-read the saved AI engineering articles, add your own commentary, tag them (e.g., harness-eng), and move them into tech/AI. ## Quick Start Run the weekly review for my Inkdrop inbox covering 2026-04-28 through 2026-05-04 and help me triage the read-later articles.